Combined geometric and texture features for face recognition
Amal Seralkhatem Osman Ali, Vijanth Sagayan Asirvadam, Aamir Saeed Malik, Azrina Bt Abd Aziz, Sarat C. Dass · 2015
Up to the present moment, there is lacking a suitable database for the purpose of research into the field of facial recognition to combat terrorism and acts of violence. As far as we know, there are no databases in existence that is publicly available and specifically incorporates photographic images of fugitive criminals and dangerous individuals. With aim to aid the afore-mentioned research effort, this work sets out to present a solution with the formation of a visual database of faces known as "Faces of the most wanted list". The current work's database model is expected to demonstrate holistic variations involving lighting effects, pose, focus, resolution, facial expression, age, ethnicity, gender, accessories, make-up, photographic quality, occlusions and background. The prevailing fugitive face database model incorporates lighting effects, pose, focus, resolution, facial expression, age, ethnicity, gender, accessories, make-up, photographic quality, occlusions and background. In addition to outlining the specifics of the database, presented in this work a system that combines geometric and texture facial features is proposed and tested over our new database.