AN UNSUPERVISED LEARNING NETWORK FOR FACE IDENTIFICATION AND LOCALIZATION

摘要

The detection, localization and identification of faces in an image are some of the essential intelligent computer vision activities. The objective of face identification is to identify or recognize the faces and / or localize these faces. The entire process should preferably be almost orientation or rotation invariant. The present paper aims to employ Competitive Learning Networks to perform these face identification tasks. The present technique is partially independent of orientation or rotation of the faces in the image. Also the identification rate of the faces is quiet moderate.