An AI-Based Erythema Measurement System for Psoriasis Lesions

Psoriasis is a common chronic inflammatory skin disease. Disease severity is commonly assessed using the Psoriasis Area and Severity Index (PASI), in which erythema is graded subjectively on a 0-4 scale. This visual assessment is prone to significant inter- and intra-rater variability. Although objective tools such as colorimeters provide accurate erythema measurement, their high cost limits routine clinical use. Smartphone imaging combined with artificial intelligence (AI) offers a practical alternative for objective assessment. However, variability in lighting conditions can affect image consistency. Incorporating a color calibration card enables accurate color normalization. This study aims to develop and validate an AI-based system for measuring erythema in psoriatic lesions using smartphone images with a color card, compared against a standard colorimeter to assess validity and reliability.

Inclusion Criteria: * Adults aged ≥18 years * Diagnosed with psoriasis by a dermatologist * Presence of at least five clearly photographable lesions * Willing and able to take photographs using a smartphone with both a standard color calibration card and a study-specific color card at home weekly, and submit images via Line or email * Willing to have clinical photographs taken at baseline and at a follow-up visit * Provides written informed consent to participate in the study Exclusion Criteria: * Presence of skin infection or other lesions that may interfere with color assessment (e.g., tattoos) * Use of self-tanning products or substances that alter skin color * Inability to perform photography or comply with study procedures