Publications

Abstract : This framework deals with a distinct problem domain where face recognition is performed for individuals who have under- gone gender transformation over a period of time. The recogni- tion rate of face recognition stands a major challenge in deal- ing with the pictures or video frames of transgender individu- als. Typically the sexual orientation change causes serious modi- fications in the actual appearance of the face just as in the body of a transgender individual. Therefore, it presents extra complex- ity / burden in taking care of the accuracy as far as transsex- ual face acknowledgement. Subsequently, there is a requirement for face recognition framework to reliably distinguish the peo- ple after they go through sex change. As Convolutional Neu- ral Network (CNN) has demonstrated to be one of the powerful tool in dealing with feature extraction in images, a new frame- work is presented which uses CNN to increase the recognition rate in transgender images. The proposed model extracts the fea- tures of transgender’s face components such as two eyes, nose and mouth using CNN. The CNN have been utilized in the proposed model. The investigations were done on HRT transsexual database.