Tokyo Denki University and Albion Co., Ltd., which are conducting joint research on a new skin analysis method using deep learning, have developed a technology that uses deep learning to generate images of unsurfaced blemishes and a technology that uses deep learning to Presented research results at an academic conference on technology for predicting and imaging future skin through continuous use of cosmetics.
The first research is a technology that uses deep learning to generate images of spots that do not appear on the surface, especially for spots that gradually appear on the surface of the skin after the age of 1, without using a special camera.Through this research, we were able to discover a technique for generating images of unsurfaced blemishes from skin images taken with a normal camera.
The second research is "Technology for predicting and imaging future skin through continuous use of cosmetics using deep learning."Focusing on the problem that it is extremely difficult to predict the future of skin that has continued to use cosmetics from the current skin condition, we applied deep learning to images of skin that has continued to use cosmetics for a certain period of time. We investigated how to predict the effects of continued use.Based on this research, we constructed an algorithm that predicts and visualizes future skin with continuous use of cosmetics from current skin images.
By utilizing the results of this research, it will be possible to confirm the current skin condition, predict the future skin, and propose a skin image after continuing skin care.In the future, we aim to utilize this technology as a help in choosing appropriate products and increasing motivation for skin care.In the future, we will further improve the accuracy of the skin analysis method and proceed with systematization.
*Deep learning is a method that uses a model called a neural network that reproduces the mechanism of human nerve cells to make a computer learn in the same way that humans naturally perform cognition and learning, such as image recognition and voice recognition. One.