Amorepacific Develops "AI-Based Automatic Skin Irritation Diagnosis Technology"

Enhancing AI Research Productivity
Strengthening Global Leadership in Safety Assessment

On September 10, Amorepacific announced that it has independently developed an "artificial intelligence (AI)-based automatic skin irritation diagnosis technology." The related research paper was published in the online edition of "Contact Dermatitis," the official journal of the European Society of Contact Dermatitis (ESCD).

In this study, Amorepacific developed an AI-based model that automatically diagnoses skin irritation reactions in patch test images by utilizing the latest deep learning technology, the YOLOv5x object detection algorithm.

R&I Center Mijiwoom. Provided by Amorepacific

The AI diagnostic model developed by Amorepacific was trained on a total of 83,629 skin data samples, each scored from 0 to 4 for irritation by experts. Its performance was validated using 1,312 evaluation data samples and 1,536 verification data samples. The developed AI model achieved a high accuracy rate of 98.3% at both the 24-hour and 48-hour marks, and demonstrated high reliability with a sensitivity of 99.7% for an irritation score of 0 (no irritation).

Kim Youngso, Head of the Quality and Safety Research Laboratory at the Amorepacific R&I Center, stated, "The AI-based automatic skin irritation diagnosis technology, published in an international journal, is an innovative achievement that significantly enhances efficiency while increasing the accuracy and reliability of diagnoses. Moving forward, Amorepacific will continue to advance its next-generation skin safety assessment platform by integrating AI technology, providing trusted beauty solutions to customers worldwide."

This content was produced with the assistance of AI translation services.