Nasolabial Wrinkle Segmentation Based on Nested Convolutional Neural Network

Wrinkles one of the common structures on human faces. Their detection is often challenging to effectively cope with skin images and can be an important step for many different applications. Skin wrinkle segmentation play an important role in face-feature analysis and assessing the beneficial effects of dermatological and cosmetic anti-aging treatments. Existing approaches of the image-based analysis of wrinkle extraction performance, which usually decreased because of weakness of wrinkle edges and similarity to the surrounding skin. In this paper, nested convolution neural network is applied to extract nasolabial wrinkles from facial images. In addition we applied a structure of deep encoder - decoder style network suitable for nasolabial wrinkle extraction. The proposed nested network, shows state-of-the-art results obtained an accuracy of 98.9%, which demonstrate novelness of this method