Infrared thermal imaging for assessing human perspiration and evaluating antiperspirant product efficacy | Scientific Reports

Oct 23, 2024

Scientific Reports volume 14, Article number: 24994 (2024) Cite this article

Metrics details

In humans, perspiration regulates core body temperature. Therefore, objectively evaluating it is essential for studying sweat gland function and mechanisms, particularly in antiperspirant efficacy studies. Various approaches have been developed for measuring human perspiration and evaluating antiperspirant efficacy, but are unsuitable for robust and routine clinical testing applications. This paper shows how infrared thermography, utilizing both high- and low-resolution modes, functions as a multiscale imaging modality. The high-resolution mode extracts physiological parameters (respiratory ~ 0.3 Hz and heart rate ~ 1.0 Hz) and visualizes the reduction of the sweat pore radii (from 359 ± 155 μm to 161 ± 47 μm) after antiperspirant application, consistent with known mechanisms of pore plugging and constriction induced by aluminum salts. The low-resolution mode quantitatively maps sweat retention in underarm clothing. All study participants in a clinical trial showed reduced sweat retention on their T-shirts due to antiperspirants, with reductions ranging from approximately 37–97% and an average reduction of 77.7 ± 22.1% using the developed methodology and tested antiperspirant. Overall, this non-invasive technique presents significant potential for clinical and personal care product evaluations, particularly in the early stages of product development.

Perspiration is an inherent and vital physiological process that plays a critical role in thermoregulating the human body1. This essential mechanism maintains stable internal temperatures during physical exertion or exposure to high heat. The process begins when the hypothalamus senses an increase in body temperature, activating eccrine glands to secrete a saline solution through ducts into the skin surface. As this sweat evaporates, it dissipates heat, effectively cooling the body. While eccrine glands are primarily responsible for thermoregulation, apocrine glands, located predominantly in the axillary region, contribute to body odor due to bacterial metabolism of organic molecules in their secretions2. The control of axillary odor is effectively managed through the use of deodorants and antiperspirants. Deodorants typically incorporate antibacterial agents that eliminate bacteria and hinder bacterial growth, thereby reducing odor. Conversely, antiperspirants work by temporarily impeding perspiration from reaching the skin’s surface, keeping the area dry3. They also serve as antibacterial agents, controlling underarm malodor. Because antiperspirants impede the sweating process, they are considered drug products by the U.S. Food and Drug Administration (FDA) and are regulated to only being applied to the underarm4. Deodorants are considered cosmetic products and are not limited to where they can be applied. The global deodorant market, driven by the growing global population’s increased awareness of the importance of personal hygiene, was valued at USD 22.54 billion in 2018 and is projected to reach USD 30.76 billion by 20265.

Antiperspirants function by temporarily blocking sweat secretion from reaching the skin surface. They achieve this primarily through aluminum-based compounds that form temporary plugs within the sweat ducts, thus preventing sweat from exiting the body. This mechanism not only reduces wetness but also offers an antibacterial action, limiting odor production. Among FDA-approved antiperspirants, those containing aluminum compounds are the most prevalent6. Aluminum chlorohydrate, aluminum zirconium, and aluminum chloride are among the commonly used aluminum compounds7. To assess the effectiveness of antiperspirant products, clinical evaluations often follow the FDA guidelines, which employs gravimetric analysis for the testing of these over-the-counter products8. Briefly, this approach involves collecting sweat by placing a cotton pad beneath the subject’s axilla or back in a controlled hot environment, and the efficacy of the product is determined by gravimetric comparison with a baseline measurement. The baseline is collected after three weeks of non-antiperspirant use in order to ensure complete removal of the aluminum active ingredient. However, this gravimetric approach fails to fully replicate real-life conditions involving liquid moisture absorption and transfer, such as during physical activity or everyday tasks while wearing clothing. Secondly, it does not capture the wetness protection benefit delivered by other modes of action beyond the plugging of sweat ducts. Lastly, this method does not provide insights into individual sweat pore activation or the underlying physical interaction mechanism between sweat pores and the active ingredients in antiperspirants9. Sweat pore activation is fundamentally linked to the body’s thermoregulatory processes, which play a crucial role in maintaining a stable internal body temperature. When the body temperature rises due to physical exertion or environmental heat, the brain initiates the sweating process to dissipate excess heat and cool the body. This cooling effect is achieved through the evaporation of sweat from the skin surface, which helps to regulate body temperature. Therefore, understanding the plugging process and investigating the mode of action of various types of new active ingredients in antiperspirant research and development requires a more comprehensive understanding of these interactions. For these reasons, alternative methods are needed to explore individual sweat pore activation and the dynamics of interaction between sweat pores and antiperspirant ingredients in realistic scenarios.

In order to explore the efficacy of antiperspirants and study sweat duct physiology and human perspiration, various techniques and methods have been developed over the years. These include traditional approaches such as thermoregulatory sweat testing (TST)10, quantitative sudomotor axon reflex testing (QSART)11, silicone impressions, starch-iodine paper, plastic imprint, sympathetic skin response (SSR)12, quantitative direct and indirect axon reflex testing (QDIRT)13. Recently, optical methods such as direct and indirect video imaging14,15, infrared thermal imaging16, colorimetric mapping based on hydrochromic polymers17, optical coherence tomography (OCT)18, and multiphoton coherent anti-Stokes Raman scattering microscopy (MPM-CARS)19have also emerged. While these methods offer insights into various morphological and functional aspects of human perspiration and sweat duct activities, they are not suitable for robust and routine clinical testing applications due to their limitations. Optical techniques like OCT and MPM-CARS are time-consuming, require specialized equipment, have limited field of view, are sensitive to motion artifacts, and involve computationally intensive 3D image reconstruction and analysis methods. Moreover, these methods do not provide essential parameters such as sweat production and sweat pore activation, which are crucial for evaluating the efficacy of antiperspirant products. Preliminary work by Raccuglia et al. and Krzywicki et al. suggest that infrared thermography (IRT) can address these challenges due to its easily adjustable field of view and fast image acquisition capabilities20,21. Building on these findings, the present study focuses on the objective examination of sweat gland function and physiological parameters, particularly in antiperspirant efficacy and clinical research.

In this paper, we propose the use of IRT as a novel non-contact multiscale imaging modality for studying human perspiration and evaluating the effectiveness of antiperspirant products. We introduce two approaches for utilizing IRT to assess antiperspirant efficacy. The first approach involves high resolution thermal imaging to directly visualize sweat pores and pore activation on in vivo human skin. By employing a novel spatio-temporal analysis method, we can evaluate product efficacy based on the underlying plugging mechanism and observe individual sweat pore behavior. From the sweat pore dynamics, physiological parameters such as respiration rate, heart rate, and pore activation area diameter were extracted. The second approach employs low resolution thermal imaging and image analysis to quantitatively map sweat retention in the underarm area of clothing, providing insights into sweat production in real-life scenarios involving antiperspirant application and physical activity. A randomized controlled clinical trial involving six participants (three men and three women) was carried out to further corroborate this approach. Through these two approaches, we demonstrate the feasibility of infrared thermal imaging as a promising technique for studying the mode of action and efficacy of antiperspirants on human subjects. This non-invasive and versatile imaging modality has the potential to enhance our understanding of antiperspirant product performance and contribute to the development of improved formulations.

In infrared thermal imaging, a sensor converts infrared (IR) radiation, which typically falls within the wavelength range of 0.9–14 μm, into a visual image representing the thermal profile/ map of an object or scene. This IR or thermal radiation is imperceptible to the human eye and all objects with temperatures above absolute zero (> -273 °C) emit it as electromagnetic waves in all directions at the speed of light. The intensity of the emitted radiation increases proportionally to the fourth power of temperature, and this relationship can be described by the Stefan-Boltzmann law22,23:

In Eq. (1), E represents the energy radiated per unit area per unit time (W/\(\:{m}^{2}),\) and σ is the Stefan-Boltzmann constant (\(\:\sigma\:=5.67\times\:{10}^{-8}W.{m}^{-2}{K}^{-4}\)). This equation demonstrates that the amount of energy emitted by an object in the form of electromagnetic radiation is directly proportional to the fourth power of its absolute temperature, T (K). It is important to note that Eq. (1) applies to the ideal case of a perfect blackbody, which is an object that absorbs all incident radiation and emits 100% of the absorbed energy. In reality, objects emit radiation at a fraction of that emitted by a blackbody. The radiative properties of an object or material are expressed in terms of its emissivity (ε)24, which is the ratio of the energy emitted by an object to the energy emitted by a blackbody at the same temperature. Kirchhoff’s Law states that for any uniform medium in thermal equilibrium, the emissivity for any spectral range equals the ratio of the medium’s surface emissive power to the emissive power of a blackbody. Hence, Eq. (1) can be modified to include emissivity:

Emissivity is a real number ranging between 0 and 1, and depends on the type of material. Most objects exhibit selective radiation with emissivity strongly varying with wavelength. A high emissivity value indicates favorable radiative properties with a blackbody possessing perfect emissivity (ε = 1). Metals, such as aluminum (ε ~ 0.05–0.07), generally have low emissivity values, whereas water (ε = 0.98) and human skin (ε = 0.98) are high emissivity materials24.

For a blackbody, its spectral emission is described using Planck’s law and can be expressed as25:

In Eq. (3), \(\:{S}_{\lambda\:\:}\) represents the spectral radiance per unit wavelength (\(\:W.{sr}^{-1}.{m}^{-3})\), λ is the wavelength (m), h is Planck’s constant (\(\:h=6.63\times\:{10}^{-34}J.s\)), c is the speed of light in vacuum (\(\:c=3\times\:{10}^{8}m/s\)), and \(\:{k}_{b}\) is the Boltzmann constant (\(\:{k}_{b}\)=1.38\(\:\times\:{10}^{-23}J.{K}^{-1}\)). Mathematical plots of Planck’s law as a function of wavelength and at various absolute temperatures are displayed in Fig. 1. According to Planck’s law, human skin emits radiation ranging from ~ 8–12 μm (highlighted in red), making this an optimal window for diagnostic imaging. We investigated the thermal contrast produced by evaporative cooling in this window for perspiration imaging applications26,27. This contrast enables us to quantify sweat retention in fabrics and monitor and measure sweat pore activity directly on human subjects. When water or sweat evaporates at the skin surface - environment interface, the temperature at each point on the surface tends toward an equilibrium state. This state occurs when the local loss of latent heat through evaporation is balanced by the net heat supplied through processes such as conduction, convection, and radiation. As a result, the water cools adiabatically. Evaporative cooling is generally approximated as an adiabatic process. The evaporation of water or sweat from a fabric or sweat pores on the skin involves evaporative cooling, and thus, the temperature differences between the area exposed to sweat and its surroundings can serve as a reliable estimator for quantifying sweat retention and studying sweat pore activation.

Blackbody Radiation Simulation. Mathematical simulation of the spectral radiance of blackbody radiation at various absolute temperatures, T. The working principle of the simulation is based on Eq. (3) with T = 0–500 K25.

Commercial antiperspirant roll-on and stick products were evaluated. The roll-on format (Sanex Dermo Extra Control, Colgate-Palmolive Co.) was utilized in the high-resolution mode studies. The stick format (Clinical Strength Speed Stick, Colgate-Palmolive Co.) was utilized in the low-resolution mode studies. Furthermore, a stick placebo, containing no aluminum salt, was included in the low-resolution mode tests.

The thermal imaging experiments were conducted using two distinct setups (high- and low-resolution modes). The custom-developed experimental imaging setups were based on the FLIR A6750sc camera (FLIR System, Inc., CA). This particular camera is a Mid-Wave IR cooled scientific camera equipped with an InSb detector, which operates within the 3–5 μm waveband. The camera has an image resolution of 640 × 512 pixels with pixel pitch of 15 μm, a set frame rate of approximately 125 frames per second (fps), and an integration time of 0.53 ms. Notably, this camera demonstrates a temperature resolution of approximately 20 mK. Throughout the experiments, the factory-calibrated temperature range of the camera spanned from 10 to 90 °C. To account for emissivity considerations, the software settings were adjusted to maintain a value of 0.98.

For the controlled thermal imaging environment, experiments were conducted within a dedicated room with a consistent temperature range of 20 to 23 °C and relative humidity (rH) between 30 and 60%. To eliminate potential interference from stray or reflected background radiation, the room was meticulously sealed, and all reflective surfaces, such as glass, mirrors, and metallic objects, were carefully excluded from the experimental space. Additionally, external radiation sources, such as halogen, quartz, and tungsten lights, were effectively isolated to maintain experimental accuracy. To verify the accuracy of the camera’s calibration, Fluke 4180 series precision infrared calibrators were employed. Despite exhibiting a relative accuracy of approximately ± 2 °C, the camera offered exceptional thermal sensitivity, with a range of ± 0.02 °C. This heightened sensitivity proved to be of critical significance for our specific application, as it enabled the precise measurement of minute spatial and temporal variations in the temperature of the objects being studied.

The imaging mode shown in Fig. 2 is a high resolution configuration, which facilitates direct imaging of individual sweat pore activity in human subjects, enabling the observation of physiological behaviors. This mode was instrumental for fundamental studies of sweat pore activation and evaluating the impact of commercially available antiperspirants on sweat production.

To directly image the forehead and dorsal finger skin area, the field of view (FOV) was modified using a lens spacer (6.35 mm) in between the lens and the imaging sensor, and altering the working distance of the 50 mm fixed lens in our IRT camera. This allowed the lens (f/2.5) to focus on the subject’s forehead at a distance of ~ 180 mm with a magnified field FOV of ~ 33 mm (H) × 41 mm (V) (512 × 640 pixels). The spatial resolution of the imaging setup was ~ 130 μm (Nyquist resolution with pixel size of 65 μm) and which was optimized for visualizing sweat droplets from individual and clustered pores forming a sweat film on the human forehead. Human sweat pores typically measure 60–80 μm in diameter, with sweat droplets ranging from ~ 60 μm to 600 μm, depending on whether they form as single droplets with partial to full pore activation or coalesce into larger clusters or films. The current spatial resolution is limited by the infrared optics, but can be further increased by using high numerical aperture Mid-Wave Infrared optics.

We employed the high-resolution infrared thermal imaging mode to visualize sweat pores on the dorsal index finger skin area of a human subject before and after physical activity. This imaging mode allowed for the direct observation and quantification of sweat pore activation in response to the thermoregulatory mechanisms triggered during physical exercise, establishing a link between sweat pore activation and physiological parameters. The aerobic exercise consisted of the subject running on a treadmill at ~ 4 mph for 10–15 min. Immediately after the physical activity, the index finger of the subject was fixed under the camera and the dorsal skin area was captured over a 2 min period. This high-resolution imaging approach allowed for the precise recording of sweat pore activation dynamics during the immediate post-exercise period.

To visualize the impact aluminum salts, have on sweat pore activity, we conducted in vivo high-resolution imaging studies on a total of three subjects (three male) of varying ages and ethnicities. A commercial antiperspirant product was selected for analysis (see Materials). To ensure accurate comparison of control and treatment site and for the feature-based frame registration during image processing, a cross-shaped piece of black medical athletic tape was strategically placed in the middle of each subject’s forehead before the application of antiperspirant products. This marker served to delineate regions where the product was applied and non-product control regions, ensuring consistency and facilitating post-processing analysis.

In order to induce sweat pore activity, each subject participated in a 10 min treadmill exercise session at a speed of ~ 3.5-4.0 mph. Following the exercise, a 2 min video was recorded for each subject without the application of the antiperspirant product. This baseline video allowed for the capture of the subject’s natural sweat pore activity before the application of antiperspirant. Subsequently, the same exercise and imaging routine was repeated for each subject following application of antiperspirant products. The products were consistently applied to the right side of the cross-shaped marker 2 h before exercise. This standardized application approach ensures that the products have sufficient time to diffuse within the sweat duct. Similarly, after each exercise session, a 2 min video was recorded using the infrared camera. This post-exercise imaging session demonstrated the effects of the antiperspirant products on sweat pore activity in real-time. By conducting this systematic imaging protocol with multiple subjects, it was possible to quantitatively evaluate the impact of an antiperspirant product in reducing sweat pore activation.

High resolution imaging mode setup. Schematic of the high-resolution imaging setup utilized to study the sweat pore activation. According to the representation, the infrared thermal camera is positioned on the subject’s (a) index finger via a finger stand to image the dorsal skin area and (b) head via the head stabilizing stage to image the forehead. The digital readout is stored and initially processed using the ResearchIR software (FLIR System, Inc., CA) on the connected PC. Figure 2 was created using Microsoft PowerPoint.

To quantitatively characterize sweat pore activation dynamics, a multi-step spatio-temporal image processing algorithm was developed (Fig. 3). A 2 min long thermal image sequence of the dorsal skin surface on the subject’s index finger using the FLIR A6750sc camera operating at ~ 125 fps frame rate and 0.53 ms integration time per frame was captured. Each thermal image sequence contained ~ 15,000 frames spanning the 2 min activity and recovery periods. To correct for involuntary sub-pixel subject motion artifacts during image capture, we developed an image stabilization technique based on a point feature matching algorithm implemented in MATLAB. This algorithm typically registers all frames to the initial frame, with exceptions made for system disconnections or extreme positioning differences. The image stabilization process utilizes the ‘imregconfig’ function to set up optimizer and metric configurations, using ‘OnePlusOneEvolutionary’ as the optimizer and ‘MattesMutualInformation’ as the metric. To ensure stable results, we increased the maximum iterations to 300 and reduced the optimizer’s initial radius to 1/3.5 of its original value, though this extends processing time. The transformation matrix for each frame is obtained using ‘imregtform’ with the ‘similarity’ argument, applying a nonreflective ‘similarity’ transformation that accounts for translation, rotation, and scale. We then use the ‘imwarp’ function to register the frames based on the acquired geometric transformation matrix. This workflow effectively stabilizes the involuntary motion artifacts in human finger/forehead thermal image sequences, optimizing the process for correcting movements while maintaining image quality. To maximize thermal contrast and enhance the visibility of pore regions activated at marginally higher temperatures, we applied a contrast-limited adaptive histogram equalization algorithm to each registered frame. This comprehensive approach allows for precise quantification of sweat pore activation dynamics, providing a robust foundation for further analysis in thermal imaging studies.

Following motion correction and contrast enhancement, pore regions of interest (ROI) were identified for further spatio-temporal analysis using a minimum intensity projection across the image stack to highlight areas of frequent activation. Firstly, the ROI containing visible pores were isolated from the contrast-enhanced image frames. To further segment and quantify individual pores and their activation areas in the projection map, a multilevel thresholding approach using four levels of quantization was implemented. Among the quantized levels, the darkest level was determined to provide optimal delineation of pore boundaries for segmentation. This level selected pixels with the highest frequency of activation over the sequence, allowing accurate identification and distinguishing of activated pores. Next, a watershed transform algorithm was employed to label and separate the activated pores in the quantized projection map. To calculate the sweat pore area, the image was first converted to grayscale and a specific ROI was cropped. Adaptive thresholding was then applied to generate a binary image that distinguishes foreground objects (pores) from the background. To ensure accurate detection of the foreground, the image was inverted. The area of these detected pores was measured and converted to square millimeters based on the resolution of the image. This process was repeated for approximately 80 pores to quantitatively evaluate sweat pore activation throughout the image sequence.

By consolidating all activated pores from the sequence into a single projection map, the pores that were activated at any point during the imaging period were effectively highlighted. This enabled dynamic tracking of pore behavior over time in response to induced activity and recovery. Within specified pore ROI, Fourier domain analysis was conducted to characterize changes in activation frequency and duration over the recording period in response to physical activity levels and due to antiperspirant application.

Automated image processing workflow. To extract physiological information, the pore activation dynamics data (refer to Fig. 2 for the imaging setup) was inputted into the developed workflow. The acquired image sequence is processed to stabilize the image and an adaptive histogram equalization is performed. From the post-processed data, respiration and heart rates are extracted by spatio-temporal and Fourier domain analysis and quantification of pore activation by minimum intensity projection and thresholding.

The low-resolution configuration was utilized to visualize the entire underarm area of a t-shirt. A 50 mm lens without any spacers was used and the FOV was adjusted to cover an area of approximately 780 mm (horizontal) × 615 mm (vertical) with a spatial resolution of approximately 2.5 mm. This spatial resolution was carefully optimized to enable precise quantification of sweat retention in the underarm areas of the T-shirts used in the study. Figure 4 depicts the schematics of the experimental imaging setup utilized during the clinical testing. This setup allowed us to capture sweat retention patterns and corresponding temperature profiles across the entire T-shirt fabric. By evaluating reductions in sweat retention and associated temperature profile changes, this imaging setup facilitated the assessment of antiperspirant efficacy in the clinical testing studies.

Low resolution imaging mode setup. Schematic of the low resolution imaging setup leveraged for the assessment of antiperspirant efficacy in the clinical studies. According to the representation, the T-shirt is fixed onto the fabric wall and the infrared thermal camera is projected onto the T-shirt/ fabric wall assembly. The digital readout is stored and initially processed using the ResearchIR software (FLIR System, Inc., CA) on the connected PC. Figure 4 was created using Microsoft PowerPoint.

To assess product efficacy at reducing perspiration for up to 96 h post application, clinical studies were performed using the low-resolution imaging mode. A commercial antiperspirant and a placebo product were utilized in this analysis (see Materials). A total of six subjects (three male and three female) were recruited to participate in the study. Prior to the product testing phase, the panelists were instructed not to use any antiperspirants, deodorants, talcum powder, or other products in their underarms for 21 days. Shaving was allowed at least two days before the study began, but no further shaving was permitted during the 12-day study period. Additionally, the subjects were not allowed to go swimming during the study period.

On Day 1, the subjects reported to the testing facility for baseline testing. An on-site medical professional took their vital signs, after which the panelists changed into the provided 100% cotton T-shirts. They then participated in a treadmill exercise session, running at approximately 6 mph for about 15–25 min in 5 min intervals. Staff present during the study visually evaluated the underarm area of the T-shirts for sweat stains every 5 min. Once the sweat in the underarm area became noticeable, running was stopped, and the subjects were asked to remove the T-shirts and hand them over to the staff for baseline thermal imaging. From Day 1 to Day 8, the subjects washed their underarms with a provided Camay Classic bar soap (Unilever International, Inc.), and the test products were applied once daily according to the study randomization (highlighted in the use instructions). Each subject applied five strokes of each testing product, delivering approximately 0.5 g of the product, and ensuring uniform coverage of the center area of the axillary vault (approximately 4 × 6 in). After the last daily application on Day 8, the panelists were instructed not to wash, wet, shave, or apply any products to their underarms until Day 12. On Day 12, the subjects returned to the testing facility to repeat the same study procedures as on Day 1. The subjects removed their T-shirts and provided them to the study staff to hang vertically using plastic hangers. Imaging was performed immediately after the T-shirts were hung to prevent distorting the retained sweat. The schematic representation of the clinical testing protocol is illustrated in Fig. 5, showcasing the step-by-step process followed during the study.

Step-by-step procedure for the human-use study. The schematic shows that the recruited subjects undergo a mandatory 21 day washout period. Next, baseline imaging is performed on the recruited subjects (red box, left panel). Afterwards, 8 days of product application is performed following the randomization procedure. Finally, 4 days (96 h) are allowed to elapse before the last phase of imaging (red box, right panel). Figure 5 was created using Microsoft PowerPoint.

Three male and three female subjects were recruited for the study. Prior to inclusion of the study, subjects will be asked to use only the test products provided on the appropriate underarm as per the study randomization.

Inclusion criteria for the healthy volunteers in the study were as follows: (1) males and female, age 18 through 60, in general good health; (2) subjects must be willing to participate as evidenced by reading and signing the Informed Consent Form; (3) subjects who are willing to have vitals (temperature, heart rate, & blood pressure) taken by a medical professional during both study visits; (4) subjects who confirm that their vital signs are within the acceptable range: (1) Blood pressure 140/90 mmHg or below; (2) Temperature 99.2 °F or below; (3) Pulse 100 bpm or below; (5) subjects who are willing to refrain from using talcum powder, antiperspirants or deodorants or any other products under his or her arms for 21 days and only use the products given in the underarm area during the study; (6) subjects who have shaved his or her underarm area at least 2 days prior to the study and are willing to refrain from shaving during the study; (7) subjects who confirm to the best of their knowledge that they are physically able to complete aerobic exercises that involve running at approximately 6 mph on a treadmill for approximately 15 to 25 min; (8) subjects who are willing to complete running at approximately 6 mph on a treadmill for at approximately 15 to 25 min under ambient conditions (outdoors).

Immediately after the aerobic exercise, the T-shirts of the subjects were collected and fixed onto a fabric wall and a set of images of both front and back side were collected. Image acquisition was performed in the imaging lab, which has a temperature of approximately 77 °F and relative humidity around 30% rH. To improve the signal-to-noise ratio, 10 thermal images of both the front and back side of the T-shirts with the collected sweat were recorded. For quantitatively analyzing the sweat retention in the underarm area, the acquired images of both front and back sides of the T-shirt were averaged separately. Then, a semi-automated approach was used to segment the underarm areas with retained sweat. The areas of interest were manually selected on a graphical user-interface, which automatically segments the ROI using the region growing algorithm. Next, the sweat retention was estimated by separately adding the weighted area intensity of pixels segmented under the left and right underarm for both the front and back sides. Based on Fig. 6, the sweat retention in the left (SRL) and right (SRR) regions of the underarm areas were estimated as SRL = SRLF + SRLB and SRR = SRRF + SRRB where SRLF, SRLB, SRRF, and SRRB are the sweat retention estimated separately for each of the underarm areas of left-front, left-back, right-front, and right-back, respectively. PNand IMean are the number of pixels and the mean intensity of the corresponding segmented underarm sweat retention areas. Finally, the ratio of sweat retention (µ) was calculated using the following equation:

Where SRC and SRE are the sweat retention of the control and experimental sites respectively, and these sites can be either SRL or SRR or vice versa. Equation (4) was used to normalize the acquired data by converting the weighted area intensity of the pixels into percentages. These calculations rely on setting one underarm to 100% and the value for the opposite underarm to 100 ± µ. Thus, an underarm that contains 339,280 pixels on the right-side and 302,656 pixels on the left-side was normalized by fixing one side as the reference while scaling the other side’s pixel count. For example, fixing the right side at 100% with 339,280 pixels, the left side with 302,656 pixels corresponds to 89.2% of the right side (µ = 10.8). Alternatively, fixing the left side at 100% with 302,656 pixels, the right side at 339,280 pixels is 112.1% of the left side (µ = 12.1). This normalization process allowed quantitative comparison between the two unequally-sized underarm sides. Representing the retained sweat as percentages allows the changes in the weighted area intensity of pixels to be accurately calculated since the amount of sweat naturally varies day-to-day in human subjects. The full calculation and corresponding data is detailed in the supplementary information.

Quantifying the retained sweat from the human-use study. The sweat quantification approach is demonstrated using a representative image (refer to Fig. 4 for the imaging setup). The denoted right (R) and left (L) symbols are assigned based on the reverse images, i.e. the subject facing forward. The sweat regions in the underarm (black spots on the T-shirt) are segmented using a custom-built MATLAB algorithm. The segmented areas were enlarged, background subtracted, and vertically offset from the original image for clarification.

In this study, the high-resolution imaging mode was employed to visualize sweat pores on the dorsal finger skin area of a human subject before and after physical activity. The thermal images allowed for the direct observation and quantification of the activation of sweat pores in response to thermoregulatory mechanisms triggered during physical exercise.

The research described herein focused only on systematically studying sweat pore dynamics induced by exercise. Under normal room conditions (23 °C and 30% rH), the subject’s dorsal skin area of the index finger did not exhibit significant sweat pore activation. The RGB image of the dorsal skin surface for the human subject before physical activity is captured in Fig. 7A and the corresponding thermal image is presented in Fig. 7B. The same area was imaged after physical activity (Fig. 7C). An increased amount of active sweat pores was distinctly observed, as reflected by the black dots in the image, due to the sweat secretion. To further demonstrate that the presence of the black spots were due to sweating, we performed optical coherence tomography (OCT) studies. Due to the interference of melanin when imaging the dorsal skin with OCT, the palmar side of the same finger was analyzed. Supplementary Fig. 1 shows that when the sweat pores are highly active, sweat secretion from the finger can be clearly detected.

Visualization of sweat pore activity. (A) RGB and (B) Thermal image of the dorsal skin area of a human subject before physical activity. (C) Thermal images of the dorsal skin area of a human subject after physical activity (Media 1). The ROI is outlined in the black box. Increased amount of active sweat pores (black dots in the image) can be seen after physical activity, C. Representative images are shown. Scale bars, 5 mm.

To gain a deeper understanding of the spatio-temporal activity of sweat pores, thermal images of the dorsal skin area of the subject’s index finger over a 2 min period were captured. A temporal analysis on selected sweat pores were performed to investigate their behavior in response to physical activity (Fig. 8A and B). This enabled temporally resolved images of the sweat pore dynamics to be created. A distinct pattern of periodic activation and deactivation can be observed (Fig. 8C). To elucidate the relationship between sweat pore dynamics and underlying physiology, a fast Fourier transform was applied to the thermal intensity profile of the selected sweat pore. The resulting frequency power spectrum (Fig. 8D) reveals distinct frequency components matching known physiological parameters. Peaks at ~ 0.3 Hz and ~ 1.0 Hz correspond to respiration and heart rates, respectively, indicating that sweat pore activation is synchronized to these rhythms. These observed features also agree with previous reports for where these modes should appear28,29. Furthermore, analysis of the same skin region, without physical activity, shows an absence of these characteristic peaks (Supplementary Fig. 2).

In order to quantitatively assess sweat pore activation throughout the image sequence, a series of image processing steps were completed (Fig. 8E and F). To visualize the segmentation results, we assigned randomized color labels to each segmented activated pore area (Fig. 8G). In order to mitigate over-segmentation, the watershed algorithm utilized a city-block distance metric and median pre-filtering for noise reduction. Following watershed segmentation in the selected FOV, we quantified approximately 80 segmented pores to generate a pore size distribution (Fig. 8H). The analysis revealed activated pore and corresponding sweat droplets (single and coalesce into larger clusters or films) diameters ranging from 50 to 325 μm, with a mean pore activation area diameter of 216 ± 72 μm. These results demonstrate the direct link between sweat pore behavior and cardiovascular/respiratory activity during physical exertion. The periodic activation of sweat pores serves as a surrogate indicator of thermoregulatory and circulatory processes in response to exercise-induced heat stress. High-speed infrared imaging provides a non-invasive means to sensitively capture this interconnectivity. Moreover, these results highlight the wealth of physiologic information encoded in sweat pore activation dynamics measurable with a high-speed thermal video.

Extracting physiological information from the sweat pore dynamics. (A) Thermal imaging stack of the dorsal skin area of a human subject after physical activity. The first and last image in the stack are f1 and fn, respectively. (B) The acquired image sequence is processed to stabilize the image and an adaptive histogram equalization is performed. The ROI is outlined in the black box. The controlled volume is projected from f1 to fn to calculate the sweat pore dynamics. (C-D) When the ROI is focused on an individual sweat pore, respiratory (lower frequency mode, ~ 0.3 Hz) and heart rate (higher frequency mode, ~ 1.0 Hz) information are extracted by spatio-temporal and fourier domain analysis on the controlled volume. (E-H) When the ROI is focused over a distribution of sweat pores, pore activation quantification can be extracted by minimum intensity projection and thresholding. The activated pore area diameters range from 50 to 325 μm, with a mean pore activation area diameter of 216 ± 72 μm. This is a demonstration of the workflow described in Fig. 3.

Furthermore, high resolution imaging studies were conducted to demonstrate the impact antiperspirants have on sweat pore activity (n = 3, Fig. 9A). We opted to use the forehead for testing the antiperspirants instead of the finger, as demonstrated above, due to its larger surface area for product application. Prior to imaging, participants were instructed not to use any facial products that could potentially interfere with the natural behavior of sweat pores in the forehead region. As illustrated in Supplementary Fig. 3, each subject’s forehead was divided into two sections - a control site and an experimental site - for baseline and product testing. During the baseline test, both the control and experimental sites were product-free. In the product test, the control site remained product-free while the experimental site was treated with the aluminum-based antiperspirant. Figure 9B depicts the quantitative maps of the sweat pore distribution for each subject at baseline without physical activation (i, ii), after activation (iii, iv), and after product application to the experimental site of the forehead (v, vi). On the experimental site, the sweat pore radii decreased from 359 ± 155 to 161 ± 47 μm after antiperspirant application, with the control site approximately demonstrating the same level of activity and pore radii.

From these pore distribution maps, the estimated total pore activation area was then calculated. For Subject 1, the pore activation areas at baseline (no activation) were 2.46 mm2 and 0.75 mm2 for the control and experimental sites, respectively as shown in Fig. 9C (i). After physical exercise to stimulate sweating, the baseline pore activation area increased to approximately 90 mm2 for both sites. The antiperspirant product was then applied only to the experimental site. Interestingly, while the untreated control site maintained a high level of pore activation (around 80 mm2), the product-treated experimental site showed dramatically reduced pore activation (0.75 mm2). This response was consistent across all subjects as shown in Fig. 9C (ii), indicating that antiperspirant application suppresses sweat pore activity by reducing pore radii and activation area. Through this robust processing and analysis workflow, we gained valuable insights into pore physiology, allowing us to understand the activation patterns and characteristics of sweat pores. Moreover, this methodology holds promise for evaluating the effectiveness of antiperspirant treatments targeting sweat activation. By objectively quantifying pore activation of the eccrine gland in response to antiperspirant applications, the efficacy of different antiperspirant formulations can be quantified early in the research and development process.

Influence of antiperspirants on sweat pore dynamics. (A) Processing approach for the acquired forehead images. Thermal imaging stack of the human subject at (i) baseline, (ii) after physical activity, and (iii) after product application to the experimental site (Media 2). Motion correction (Media 3) and minimum intensity projection are performed on the corresponding raw data points of (i), (ii), and (iii) to produce (iv), (v), and (vi). (B) Control (first column) and experimental (second column) sites (i, ii) at baseline (iii, iv) after physical activity, and (v, vi) after product application to the experimental site. Binary and pore size segmentation were used to approximate the sweat pore radius. The calculated pore radii range from 100 to 1000 μm, with some pores sporadically appearing at the baseline (mean pore radii at control/experimental sites ~ 247 μm). The sweat pore radii decrease from 359 ± 155 to 161 ± 47 μm after product testing, with the control site approximately demonstrating the same level of activity. (C) Plot of the pore activation area, highlighting the difference in the rest/ active states at baseline and the quenching response of the antiperspirant product (i). Corresponding pore activation results for each subject after product testing (ii). This is another demonstration of the workflow described in Fig. 3.

Finally, to evaluate the feasibility of applying thermal imaging to quantify sweat produced in the underarm, a randomized controlled clinical trial (n = 6) was conducted. A representative image from the clinical trial is shown in Supplementary Fig. 4. The images, processed using the MATLAB segmentation algorithm (depicted in Fig. 6), are presented in Supplementary Fig. 5. The collected images from the study were quantified (Supplementary Table 1) and presented in Fig. 10. A baseline measurement was taken (Fig. 10A) in order to understand the innate sweating behavior of the subjects. The baseline measurement was taken 21 days following no antiperspirant product underarm use. The data obtained from the baseline measurements illustrated that the production of sweat in the left and right underarm is asymmetric and a function of subtle physiological factors such as handedness of the subject, gait, etc. Interestingly, similar observations were noted by Ferdenzi et al30.. The authors demonstrated that the asymmetric production of sweat results in a favorable environment for the increased production of microorganisms on the dominant-side. For this reason, there is also an evident side-related dependence of odor due to the microflora of the axilla.

The products (placebo, P and antiperspirant, AP) were subsequently applied to the underarms of the subjects and the result of this is presented in Fig. 10B. The product application was randomized so as not to bias the study results. Thus, any observed deviation from normal sweating behavior can be directly attributed to the presence of the antiperspirant. As previously described in the Methods section, the sweat retention measured in pixels is converted into percentages, allowing for a standardized comparison of the results. Supplementary Table 2 shows the calculation for µ after baseline and product tests. As shown in Supplementary Tables 3 and 4, the data is normalized by using the placebo-treated underarm as an internal reference that is fixed to a specific value. This eliminates any inherent variability or baseline differences between subjects, so that the normalized data reflects the true effects of the active treatment relative to the placebo control, enabling more accurate comparisons across the entire study population. The resulting tabulated values are plotted in Supplementary Fig. 6 (A and B; for subjects 1, 3, and 5) and 6 (D and E; for subjects 2, 4, and 6).

To better illustrate the results of the clinical trial, Supplementary Fig. 6 C and 6 F are combined and displayed in Fig. 10C. By overlaying the normalized data on top of each other, significant changes to the production of sweat can clearly be observed in the antiperspirant-treated underarm. The percent reduction in retained sweat between the baseline and product tests are plotted in Fig. 10D. All the participants in the study showed a reduced sweat retention on their T-shirt due to the antiperspirants, ranging from ~ 37% to ~ 97%. In subjects 1, 3, and 5, where the antiperspirant was applied to the left underarm, the sweat retention decreased from 89.2 to 4.4%, 68.7–43.5%, and 11.7–3.1%, respectively. In subjects 2, 4, and 6, where the antiperspirant was applied to the right underarm, the sweat retention decreased from 339.0 to 76.3%, 486.7–14.2%, and 233.0–32.3%, respectively. We found that an average percent reduction in sweat retention of 77.7 ± 22.1% is achieved using the developed methodology and tested antiperspirant. The standard deviation is generated using the six independent subjects. This evident reduction in retained sweat is due to sweat pore inactivity that is created with the use of aluminum salts31, which is consistent with Fig. 9. Therefore, these findings demonstrate the ability of thermal imaging and image analysis to quantitatively map and systematically monitor sweat production in the underarm under real-life conditions.

Quantification of the antiperspirant clinical trial. (A, B) Raw data of the baseline and product tests for the 6 recruited subjects (3 male and 3 female). To normalize the results, the raw data is converted into percentages using the placebo-treated underarm as a control. (C) The normalized data of the baseline and product tests are overlaid. (D) The percent reduction in retained sweat between the baseline and product tests for the underarm treated with the antiperspirant. The subjects present in the study showed reduced sweat retention on their T-shirt, ranging from ~ 37% to ~ 97%, due to the antiperspirants. In B, the P refers to Placebo and AP refers to Antiperspirant. In C, the underarm with the AP is overlaid with the respective baseline to illustrate the reduction in retained sweat onto the T-shirt. The data is plotted using a logarithmic scale and the underarm with the P is omitted because there is no change between the respective tests. A representative image from the clinical trial is shown in Supplementary Fig. 4. The processed images, using the MATLAB segmentation algorithm (depicted in Fig. 6), are presented in Supplementary Fig. 5.

The high-resolution imaging of sweat pores before and after antiperspirant treatment provided direct visualization of product effects at the microscopic pore level. By thermally profiling individual pore behavior in a spatio-temporal manner, local interactions between active antiperspirant ingredients and sweat gland activity could be observed and quantified. The developed imaging and analysis workflow enabled sensitive detection of even minute changes in pore activation and diameters down to the ~ 100 μm scale in response to induced sweating and antiperspirant application. Across multiple subjects, aluminum-based antiperspirants were found to dramatically suppress pore activation areas by up to 90% compared to baseline measurements. This aligns with previous histological evidence suggesting aluminum antiperspirant actives cause pore plugging and constriction effects that inhibit sweat release32. Interestingly, the temporal analysis of pore activation dynamics revealed distinct rhythmic patterns synchronized with underlying physiological signatures. Through Fourier transform analysis, the periodic fluctuations in thermal intensity could be correlated to respiratory and cardiovascular rhythms. This further highlights the close inter-connections between thermoregulatory sweating mechanisms, pore behavior, and systemic physiology. High-speed infrared imaging elucidated these relationships in a non-invasive, highly sensitive manner unmatched by traditional staining or imprint techniques33,34,35.

By enabling direct visualization of pore activation, the imaging approach could provide further insights into the kinetics of antiperspirant interactions as well as variability across different skin regions and subject populations36. The technique’s non-invasive and non-contact nature is ideally suited for clinical translation and testing of emerging antiperspirant active ingredients that specifically target sweat pores. Ongoing developments in infrared detectors, lenses, and multivariate analytical techniques hold promise for even higher resolution pore mapping down to ~ 50 μm scales37,38,39. These results using the high-resolution mode demonstrate the ability of infrared thermography to visualize and quantify the sweat pore plugging dynamics at the microscopic level.

Complementary to the high-resolution analysis, the clinical testing performed using the low-resolution imaging mode illustrated the capability of infrared thermography for quantitatively evaluating antiperspirant performance under real-world conditions. By imaging sweat retention on fabrics, we could systematically monitor the visual impact of sweating and its mitigation after the use of an antiperspirant. The developed workflow enabled sensitive detection and quantification of minor reductions in sweat accumulating on to fabric after antiperspirant treatment. On average, the tested aluminum-based antiperspirant formulation demonstrated sweat retention reductions of around 77.7% compared to baseline measurements (pre-antiperspirant treatment). This evident reduction of sweat moisture aligns with the known mechanism of aluminum salts which drive the antiperspirant benefits. Aluminum salts form plugs within sweat ducts through interactions with keratin fibers. The obstructed pore impedes sweat secretion onto the skin surface40,41. The considerable sweat retention reductions measured through infrared imaging are consistent with these plugging effects. Clinical studies with small sample sizes may lack sufficient statistical power to detect certain responses reliably. Statistical analysis with baseline sweat as a covariate, confirmed that the reduction in sweat retention was statistically significant (p < 0.05). However, it is crucial to acknowledge that with a small sample size (n= 6), the study’s statistical power is limited, which may affect the robustness of the findings. For a more thorough assessment, future studies should include larger sample sizes to increase statistical power and confirm the reproducibility of these results. This study complies with FDA guidelines for Effective Testing of OTC Antiperspirant Drug Products. Despite the limited sample size in our clinical study, we have successfully demonstrated the efficacy of this method using a widely recognized material with a proven mode of action. The imaging approach also revealed subject-specific variability in the extent of antiperspirant-induced sweat reduction. While most subjects showed over 70% sweat reductions, one subject exhibited only around a 36.8% percent change. This points to potential differences in individual sweat pore anatomy, pore plugging susceptibility, and physiological sweat production rates that warrant further investigation41,42. Specifically, factors like baseline sweat rates and skin type can influence the effectiveness of the antiperspirant. Individuals with higher baseline sweat rates might exhibit different levels of sweat reduction due to variations in sweat gland activity and density. Variations in skin type, such as thickness and natural oil presence, may affect how aluminum salts interact with and penetrate the sweat ducts. Age is another critical factor, as it impacts both sweat gland function and skin barrier properties. Younger individuals generally have more active sweat glands, potentially influencing antiperspirant effectiveness. Moreover, differences in sweat pore size and distribution, along with physiological factors like hydration status and hormonal levels, could also significantly affect antiperspirant efficacy. Nevertheless, the technique’s sensitivity enabled clear discrimination of product performance on all subjects tested.

Notably, this infrared thermal imaging approach offers a new way of visualizing the benefits of antiperspirants. For instance, by mimicking real-life scenarios, the thermography-based method accounts for more comprehensive evaluation of sweat retention kinetics and dynamics, which is lacking in gravimetric approaches. Moreover, the non-contact, hands-free nature of infrared thermal imaging makes it better suited for clinical applications and avoids potential skin irritation or discomfort during prolonged testing. The technique’s high thermal sensitivity and fast acquisition times enable dynamic tracking of sweat evaporation on fabrics as it occurs. Raccuglia et al. identified several limitations to using IRT for monitoring sweat accumulation20. Although many of these limitations are not relevant here, sweat migration and loss continue to pose significant challenges. To mitigate this, we conduct imaging immediately following physical activity to accurately capture the retained sweat and minimize these effects. Nevertheless, further optimization of camera specifications, resolution, and quantification algorithms could enable even higher resolution sweat mapping in future studies. These results using the low-resolution mode demonstrate the use of infrared thermography to evaluate antiperspirant performance by sensitively mapping sweat retention under realistic use conditions.

In summary, IRT has promising applications in clinical imaging to study the underlying physiology of various inflammatory diseases, Raynaud’s disease, complex pain, etc. Its ability to provide objective, non-contact temperature measurements, regardless of the subject’s skin color, makes it a credible tool for understanding the thermal physiology of various clinical conditions. We have effectively demonstrated a novel imaging platform based on this technique for not only studying sweat pore activation but also quantifying sweat retention to assess the efficacy of antiperspirant products. We believe that this multiscale imaging approach has the potential to contribute to antiperspirant research and personalized product development by better simulating their real-life conditions while simultaneously providing fundamental physiological information. By leveraging the physics of thermal contrast, IRT enables us to explore perspiration imaging in a range of applications, from fabric analysis to clinical studies.

Data underlying the results presented in this paper may be obtained from the corresponding author upon reasonable request.

Baker, L. B. Physiology of sweat gland function: The roles of sweating and sweat composition in human health. Temperature (Austin) 6(3), 211–259. https://doi.org/10.1080/23328940.2019.1632145 (2019).

Article ADS PubMed Google Scholar

Natsch, A. & Emter, R. The specific biochemistry of human axilla odour formation viewed in an evolutionary context. Philos. Trans. R. Soc. Lond. B Biol. Sci. 375(1800), 20190269. https://doi.org/10.1098/rstb.2019.0269 (2020).

Article CAS PubMed PubMed Central Google Scholar

Callewaert, C., Hutapea, P., Van de Wiele, T. & Boon, N. Deodorants and antiperspirants affect the axillary bacterial community. Arch. Dermatol. Res. 306(8), 701–710. https://doi.org/10.1007/s00403-014-1487-1 (2014).

Article CAS PubMed Google Scholar

Corby-Edwards, A.K. FDA regulation of cosmetics and personal care products. Cosmetics and FDA Regulation (2013).

https://www.fortunebusinessinsights.com/enquiry/request-sample-pdf/deodorant-market-102687

Steinberg, D.C. Regulations of underarm antiperspirants and deodorants. Cosmetic Science and Technology Series 283–304 (1999)

Sanajou, S., Şahin, G. & Baydar, T. Aluminium in cosmetics and personal care products. J. Appl. Toxicol. 41(11), 1704–1718 (2021).

Article CAS PubMed Google Scholar

Maxeiner, B. et al. Design and application of a screening and training protocol for odor testers in the field of personal care products. Int. J. Cosmet. Sci. 31(3), 193–199. https://doi.org/10.1111/j.1468-2494.2009.00494.x (2009).

Article CAS PubMed Google Scholar

Teerasumran, P., Velliou, E., Bai, S. & Cai, Q. Deodorants and antiperspirants: New trends in their active agents and testing methods. Int. J. Cosmetic Sci. (2023).

Fealey, R.D., Low, P.A. & Thomas, J.E. Thermoregulatory sweating abnormalities in diabetes mellitus. Mayo Clinic Proc. 64(6) (1989).

Low, P. A., Caskey, P. E., Tuck, R. R., Fealey, R. D. & Dyck, P. J. Quantitative sudomotor axon reflex test in normal and neuropathic subjects. Ann. Neurol. 14(5), 573–580 (1983).

Article CAS PubMed Google Scholar

Vetrugno, R., Liguori, R., Cortelli, P. & Montagna, P. Sympathetic skin response: basic mechanisms and clinical applications. Clin. Autonomic research 13, 256–270 (2003).

Article Google Scholar

Gibbons, C. H., Illigens, B. M., Centi, J. & Freeman, R. QDIRT: quantitative direct and indirect test of sudomotor function. Neurology 70(24), 2299–2304 (2008).

Article PubMed Google Scholar

Gomez, D. S. et al. A new method for scoring active sweat glands. Clinics 60, 505–506 (2005).

Article PubMed Google Scholar

Lee, J. et al. Hydrochromic conjugated polymers for human sweat pore mapping. Nat. Commun. 5(1), 3736 (2014).

Article ADS CAS PubMed Google Scholar

Priego-Quesada, J. I. et al. A methodology to assess the effect of sweat on infrared thermography data after running: Preliminary study. Infrared Phys. Technol. 109, 103382 (2020).

Article Google Scholar

Park, D. H., Park, B. J. & Kim, J. M. Hydrochromic approaches to mapping human sweat pores. Accounts Chem. Res. 49(6), 1211–1222 (2016).

Article CAS Google Scholar

Kato, K. et al. Analysis of sweating by optical coherence tomography in patients with palmoplantar hyperhidrosis. J. Dermatol. 48(3), 334–343 (2021).

Article PubMed Google Scholar

Chen, X., Gasecka, P., Formanek, F., Galey, J. B. & Rigneault, H. In vivo single human sweat gland activity monitoring using coherent anti-Stokes Raman scattering and two-photon excited autofluorescence microscopy. Br. J. Dermatol. 174(4), 803–812 (2016).

Article CAS PubMed Google Scholar

Raccuglia, M., Heyde, C., Lloyd, A., Hodder, S. & Havenith, G. The use of infrared thermal imaging to measure spatial and temporal sweat retention in clothing. Int. J. Biometeorol. 63, 885–894 (2019).

Article ADS PubMed Google Scholar

Krzywicki, A. T., Berntson, G. G. & O’Kane, B. L. A non-contact technique for measuring eccrine sweat gland activity using passive thermal imaging. Int. J. Psychophysiol. 94(1), 25–34 (2014).

Article PubMed Google Scholar

Jones, B. F. A reappraisal of the use of infrared thermal image analysis in medicine. IEEE Trans. Med. Imaging 17(6), 1019–1027 (1998).

Article CAS PubMed Google Scholar

Pidwirny, M. The Nature of Radiation, Fundamentals of Physical Geography, 2nd Edition, 2006. Retrieved from http://www.physicalgeography.net/fundamentals/6f.html. Accessed in October 2023.

English, M. J. M. Physical principles of heat transfer. Curr. Anaesthesia Crit. Care 12(2), 66–71 (2001).

Article Google Scholar

Branchetti, L., Cattabriga, A. & Levrini, O. Interplay between mathematics and physics to catch the nature of a scientific breakthrough: The case of the blackbody. Phys. Rev. Phys. Educ. Res. 15(2), 020130 (2019).

Article Google Scholar

Davis, A. P., & Lettington, A. H. (1988). Principles of thermal imaging. Applications of thermal imaging, 1–34.

Chang, Y. & Wang, X. Sweat and odor in sportswear–A review. iScience (2023).

Zhang, X. et al. Contactless simultaneous breathing and heart rate detections in physical activity using IR-UWB radars. Sensors 21(16), 5503 (2021).

Article ADS PubMed PubMed Central Google Scholar

Takano, C. & Ohta, Y. Heart rate measurement based on a time-lapse image. Med. Eng. Phys. 29(8), 853–857 (2007).

Article PubMed Google Scholar

Ferdenzi, C., Schaal, B. & Craig Roberts, S. Human axillary odor: are there side-related perceptual differences?. Chem. Senses 34(7), 565–571 (2009).

Article PubMed Google Scholar

Sakhawoth, Y. et al. Real time observation of the interaction between aluminium salts and sweat under microfluidic conditions. Sci. Rep. 11(1), 6376 (2021).

Article CAS PubMed PubMed Central Google Scholar

Collins, K. J. Sweat glands: Eccrine and apocrine. In Pharmacology of the Skin I: Pharmacology of Skin Systems Autocoids in Normal and Inflamed Skin, 193–212. Springer (1989).

Shimizu, H. Shimizu’s textbook of dermatology. JA KUBU (2007).

Legner, C. et al. Sweat sensing in the smart wearables era: Towards integrative, multifunctional and body-compliant perspiration analysis. Sens. Actuators A Phys. 296, 200–221 (2019).

Article ADS CAS Google Scholar

Sharma, M., Kacker, S. & Sharma, M. A brief introduction and review on galvanic skin response. Int. J. Med. Res. Prof. 2(6), 13–17 (2016).

Google Scholar

Taylor, N. A. & Machado-Moreira, C. A. Regional variations in transepidermal water loss, eccrine sweat gland density, sweat secretion rates and electrolyte composition in resting and exercising humans. Extrem Physiol. Med. 2(1), 4. https://doi.org/10.1186/2046-7648-2-4 (2013).

Article PubMed PubMed Central Google Scholar

Sagan, V. et al. UAV-based high resolution thermal imaging for vegetation monitoring, and plant phenotyping using ICI 8640 P, FLIR Vue Pro R 640, and thermoMap cameras. Remote Sens. 11, 330. https://doi.org/10.3390/rs11030330 (2019).

Article ADS Google Scholar

Ring, E. F. J. & Ammer, K. Infrared thermal imaging in medicine. Physiol. Measur. 33(3), R33 (2012).

Article ADS CAS Google Scholar

Meola, C., et al. Infrared thermography: Recent advances and future trends (2018).

Hölzle, E. & Braun Falco, O. Structural changes in axillary eccrine glands following long-term treatment with aluminium chloride hexahydrate solution. Br. J. Dermatol. 110(4), 399–406 (1984).

Article PubMed Google Scholar

Quatrale, R., Waldman, A., Rogers, J. & Felger, C. The mechanism of antiperspirant action by aluminum salts. I. The effect of cellophane tape stripping on aluminum salt-inhibited eccrine sweat glands. J. Soc. Cos. Chem. 1981a 32, 67–73 (1981).

CAS Google Scholar

Schmidt-Rose, T. et al. Efficient sweat reduction of three different antiperspirant application forms during stress-induced sweating. Int. J. Cosmetic Sci. 35(6), 622–631 (2013).

Article CAS Google Scholar

Download references

H.M.S., T.O., and S.P. would like to thank Latonya Kilpatrick, Christina Signore, Suzie Cheng and the dermal clinical team at Colgate-Palmolive Company for conducting the preliminary thermal imaging experiments and clinical studies, respectively. A.O. acknowledges the ELO fellowship support from the Colgate-Palmolive Company. The whole research was financially supported by the Colgate-Palmolive Company.

These authors contributed equally to this work.

Colgate-Palmolive Company, Piscataway, NJ, 08854, USA

Hrebesh Molly Subhash, Tochukwu Ofoegbuna, Abmael H. Oliveira & Shyamala Pillai

Department of Biomedical Engineering, Rutgers, The State University of New Jersey, Piscataway, NJ, 08854, USA

Abmael H. Oliveira & Mark C. Pierce

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

Author contributionsH.M.S. conceived the original idea of using infrared thermography to directly evaluate the sweat pore activation and assess the efficacy of antiperspirants. H.M.S., T.O., S.P., and A.O. designed and conducted the low and high resolution imaging experiments. H.M.S., T.O., and A.O. developed the MATLAB algorithms for processing the collected images. M.P. helped with the data analysis. All authors contributed to the revision and editing of the manuscript.

Correspondence to Hrebesh Molly Subhash.

The authors declare no competing interests.

The imaging of all subjects reported in this study using infrared thermography was conducted in accordance with the guidelines of Good Clinical Practice and was approved by the US Institutional Review Board, Inc. (IRB# U.S.URB2019CP/13). All enrolled subjects signed an informed consent form.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Below is the link to the electronic supplementary material.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

Molly Subhash, H., Ofoegbuna, T., H. Oliveira, A. et al. Infrared thermal imaging for assessing human perspiration and evaluating antiperspirant product efficacy. Sci Rep 14, 24994 (2024). https://doi.org/10.1038/s41598-024-73878-8

Download citation

Received: 08 July 2024

Accepted: 23 September 2024

Published: 23 October 2024

DOI: https://doi.org/10.1038/s41598-024-73878-8

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative