We present a novel model for realizing a real-time image correction of specified objects in moving pictures. A cellular neural network incorporating human visual perception functions is proposed. It enables real-time color feature detection and smoothing interpolation of the object. This model is efficiently implemented into LSI circuits by using the finite element method. Smooth and selective correction effects are obtained in the face correction.
Takami Satonaka, Koji Asari, Shinichi Katsu, Tatsuo Otsuki, Takaaki Baba, "A Real-time Object Correction Algorithm using Cellular Neural Network for Interactive Color Image Processor" in Proc. IS&T 3rd Color and Imaging Conf., 1995, pp 126 - 129, https://doi.org/10.2352/CIC.1995.3.1.art00033