The detection of under-eye bags on the facial image

Abstract

In this paper, we propose a novel approach to detect the under-eye bags on the human facial image. The existence and exacerbation of the under-eye bags is one of the significant marks of the facial aging. Once the under-eye bags appear, it not only plays an important impact on the human looks but also gives the impression of the emaciation and senescence. Our experiments and algorithms are based on the image gallery built on our own. The image gallery consists of 8132 facial images of 230 different males which are from different age. The decision of the presence of the under-eye bags in the image is made with the help of the experienced oculists. In the analysis of the facial images, we find several significant features of the under-eye bags in the captured images and design some effective methods to detect them. Base on the our reasonable detection methods, we have designed the under-eye bags classifier. Then we use the whole gallery to evaluate the performance of the classifier. The experiments show that the under-eye bags in the facial image can be well detected with a relatively high accuracy although both the forms of the under-eye bags and the illumination of the image acquisition condition are various.