Gaussian Filter-Based Image Denoising Detects Hidden Sweat Glands and Enhances Reliability of Active Sweat Gland Density Measurements

The quantitative sudomotor axon reflex test (QSART, Iontophoresis of 10% ACh with 2 mA for 5 min) is widely used in perspiration studies. However, it faces challenges such as suboptimal image quality and misidentification of hidden sweat glands as background noise. This study aimed to apply Gaussian filtering to reveal hidden active sweat gland density (ASGD) and enhance image clarity. ASGD data of 29 participants (18 males and 11 females) were collected from eight body regions - chest, abdomen, upper back, lower back, upper arm, forearm, thigh, and calf - on both left and right sides of the body. After applying Gaussian filtering to the collected images, an improvement in image visibility was assessed. Image reliability was analyzed by evaluating the difference in ASGD between the left and right sides of the body with the Wilcoxon signed-rank test and intraclass correlation coefficient (ICC). Gaussian filtering significantly increased the visibility of sweat glands and improved ASGD detection across all eight body regions. Both symmetry and reliability of sweat gland images improved postfiltering. Applying Gaussian filtering can effectively expose previously obscured sweat glands and significantly enhance the clarity and precision of ASGD detection. Moreover, our findings identify the 0.5 × 0.5 cm² unit as the optimal measurement scale for ASGD research, which is better than the 0.25 × 0.25 cm² unit. The introduction of this advanced measurement module with superior reliability has the potential to advance active sweat gland research with QSART.