CwisarDH: a Change Detection method in Computer Vision and Video Processing based on Weightless Neural Networks

CwisarDH is a method based on Weightless Neural Networks (WNN) to address the change and motion detection problem in video processing. The main features of CwisarDH are: 1) the dynamic adaptability to background change due to the WNN model adopted; 2) the use of pixel color history buffers to improve the system behavior in videos characterized by (des)appearing of objects in the scene and slow/fast changes in lightning and background brightness and shape. CwisarDH ranked third in the CDNET 2014 challenge (www.changedetection.net) surpassing many other competitors.