Texture analysis process and analyser.

摘要

Process for analyzing textures associated with a data stream and analyzer implementing the process. The analyzer operates a transform with the aid of neural networks which are set into competition. After the solution phase the process determines that neural network delivering the best result for each class of examples presented and emits a selection index. The latter thus makes it possible to visualise the textures, for example of an image. It is possible to group the results by families. The synaptic coefficients of the neural networks can be loaded from a memory. Application to image data, acoustical data etc.