Glasses removal from facial image using recursive PCA reconstruction

Abstract

This paper proposes a new method of removing glasses from human frontal face images. We first detect regions occluded by glasses. Then we generate a naturally looking glassless facial image by recursive PCA reconstruction. The resulting image has no trace of glass frame, nor the reflection and shade made by the glasses. The experimental results show that the proposed method is an effective solution to the problem of glass occlusion, and, we believe, it can be applied to enhancing the performance of face recognition systems.