The image compression method consists in processing a current frame of pixels displaced in the image to be compressed by using a network of N neurons, 1 through N, arranged in a same plane, each neuron having a number of inputs equal to the size of frames extracted from the image and such that, after an apprenticeship phase during which the network has organized itself, and in order that the neurons represent the most probable input values, each neuron receives, at the same time, to its inputs the pixels of the current frame and outputs a signal, or potential, dj, which is positive and is all the weacker as the status of the corresponding neuron j is close to the status of the inputs ei; coding the current frame of pixels by a code associated with the neuron j of which the potential dj is minimum. The invention applies particularly to the compression of television images.
展开▼
