Skin Microstructure Segmentation and Aging Classification Using CNN-Based Modelsopen access
- Authors
- Moon, Cho-, I; Lee, Onseok
- Issue Date
- 1-Jan-2022
- Publisher
- Institute of Electrical and Electronics Engineers Inc.
- Keywords
- Skin; Aging; Microstructure; Image segmentation; Surface morphology; Deep learning; Convolutional neural networks; Skin aging; skin micro-structure segmentation; aging classification; CNN models; skin condition evaluation
- Citation
- IEEE Access, v.10, pp 4948 - 4956
- Pages
- 9
- Journal Title
- IEEE Access
- Volume
- 10
- Start Page
- 4948
- End Page
- 4956
- URI
- https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/20290
- DOI
- 10.1109/ACCESS.2021.3140031
- ISSN
- 2169-3536
- Abstract
- The skin surface is composed of a network-like microstructure comprising wrinkles. Observing and analyzing the microstructure of the skin that changes with the skin condition and aging are simple, stable, and accurate evaluation methods for skin diagnosis. However, the skin surface includes various morphological and topological changes, depending on the individual or the degree of aging. It is difficult to accurately extract and analyze a skin microstructure including these changes. Therefore, we perform skin microstructure segmentation and aging analysis by using convolutional neural network (CNN) models. First, we propose a fusion UNet model to extract the skin microstructure. We compare and evaluate the segmentation performance by using an image processing method and deep learning models. Next, we classify skin aging based on the skin microstructure. For the classification, we use four mobile CNN models: NASNet-Mobile, MobileNetV2, MobileNetV3-Small, and EfficientNet-B0. Subsequently, we compare and evaluate their classification performances. Results show that the segmentation images of the fusion U-Net are most similar to the ground truth, and the fusion U-Net model can detect fine wrinkles that are difficult to identify by the naked eye. In the microstructure-based classification of skin aging, MobileNetV3-Small exhibits the best performance with an accuracy of 94%. The proposed method facilitates an objective and quantitative analysis of the skin surface with more diverse aging characteristics. Consequently, the association between skin aging and skin microstructure changes is confirmed. Our study can be utilized in the diagnostic studies on various skin characteristics, including skin texture, anisotropy, and roughness. The proposed method can also be applied to a mobile-based self-diagnosis system.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Medical Sciences > Department of Medical IT Engineering > 1. Journal Articles
![qrcode](https://api.qrserver.com/v1/create-qr-code/?size=55x55&data=https://scholarworks.bwise.kr/sch/handle/2021.sw.sch/20290)
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.