Numerous methods have been developed for holistic face recognition with impressive performance. But Partial faces frequently appear in unconstrained scenarios, with images captured by surveillance cameras or handheld devices (e.g., mobile phones) in particular. In this paper, a method for automatically recognizing partial human face images is presented. The technique uses the Phase-Only Correlation (POC) for image matching. Experiment was conducted on a database of 479 images of 40 different persons. For experimental evaluation, a mask was generated on every query image to identify and separate the non-occluded portions. These separated portions were compared with the gallery images using the POC technique. Results have shown the proposed method is practical and provides preferable performance.
Keywords
Occlusion, Partial Face, Phase Only Correlation.
User
Font Size
Information