Soft Tissue Feature Tracking Based on Deep Matching Network

Research in the field of medical image is an important part robot to operate human organs. A intersection multi-disciplinary research fields, which direction and has achieved fruitful results. In this paper, a method soft tissue surface feature tracking based on depth matching network proposed. This described triangular algorithm. First, we construct self-made sample set for training from first N frames speckle data obtained by triangle The pre-trained ORL face then trained set. After training, carried out subsequent obtain matrix between frame. From matrix, inter-frame results can be obtained. way, completed. On basis, are compared with convolutional neural network. experimental show that proposed higher accuracy. particular, accuracy MNIST handwritten reached more than 90%.