Project Works

Retinal Sensor Array (RSA)

Digital Camera has been an important part of daily life, helping us record memorable events and scenes. In most of commercial digital cameras, Color Array Filter (CFA) has been utilized to help capture color pictures. Demosaicing algorithms are applied on CFA pictures to recover it to original full color images. In our work, inspired from human retina's configuration, we designed a new CFA called Retinal Sensor Array (RSA), which has natural anti-artifact attribution. Traditional demosaicing algorithms could not perform well on RSA images, so we proposed a new algorithm with the aid of sparse representation, multi-dictionary learning and selection. The new algorithm gain great result on RSA images compared with other algorithms.