Skin modeling and apperance reproduction considering fine hair and pigment distribution with spectral image analysis

2022 Fiscal Year Final Research Report

• Project/Area Number: 17K00255
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Perceptual information processing
• Research Institution: Osaka Electro-Communication University
• Principal Investigator: Doi Motonori 大阪電気通信大学, 情報通信工学部, 准教授 (00304155)
• Project Period (FY): 2017-04-01 – 2023-03-31
• Keywords: 画像合成 / 皮膚 / 産毛 / 色素斑

Outline of Final Research Achievements

The aim of this study is to generate photorealistic high resolution skin images. Analysis and synthesis were done for fine hair and pigmented spots, respectively. Fine hairs were observed with a microscope, and the results were used to synthesize the hair effect to skin image. The intersection point of the skin sulcus was detected by image processing, and the fine hairs were synthesized to　grow from these detected points. For pigmented spots, we created pigmented spot patterns by applying deep-learning models such as DCGAN, and synthesized them into a skin image considering the absorption of pigment based on spectral analysis.

Free Research Field

カラー画像処理

Academic Significance and Societal Importance of the Research Achievements

CGにより生成された人物画像に対する現実の人物との違和感は肌の質感が大きいと考えられる．
本研究により，肌のきめが見えるレベルの皮膚画像に対し，微細な産毛や色素の分布を考慮することで，皮膚の高質感画像生成が可能になったと考えられる．
本研究の成果は特にCGにおける質感付与に寄与できると考えられる．また，皮膚に関する医療分野での応用や化粧の効果のシミュレーションなどの産業分野での応用が期待される．