Abstract:
The results of the development of a method for recognizing and measuring objects on MRI brain inventions based on the introduction of a mathematical description of the transition area between biological objects and the subsequent decomposition of the general image of the brain into separate biological objects are presented. A distinctive feature of the developed approach is the use of neural networks only to simplify the search for key features of a given biological object, whereas the main part of the method is implemented in the form of procedures for calculating signs of transitions between biological structures by introducing a mathematical function and then filtering false detected transitions.