Stereo vision by cellular neural network with wavelet template
T. Hattori, Mamoru Tanaka · 2002
In this paper, a new method to solve the correspondence points problems for stereo vision is presented. The stereo images include surface information as a parallax. After the correspondence points are found, the parallax can be obtained. In this method, the stereo images are transformed to images mapped by a 2D Wavelet Transform by Cellular Neural Networks, and the correspondence points are found in the transferred space.