Résumé:
Image medical registration is one of the most active areas in image processing, and has attracted particular interest in many areas such as medical image analysis, remote sensing, and mapping. The main idea of these techniques is to identify the spatial shift between two images which makes it possible to match the equivalent properties. Registration can be local or international. Several deep-learning-based registration techniques have been developed that have shown impressive performance in terms of matching, which can be grouped into two main classes: Flexible and Diffeomorphic feature registration based on extraction between two images.
In the graduation project, we established an accurate learning model based on convolutional neural networks that were trained to record images. Where the model proved acceptable accuracy and speed
