作者: Bahram Gharabaghi , Guoqiang Li , Jian Wang , Jiang Tao , Ruiqi Shi
DOI: 10.1007/978-3-030-69717-4_14
关键词: Computer vision 、 Radiography 、 Image segmentation 、 Tooth root 、 Sobel operator 、 Edge detection 、 Computer science 、 Ground truth 、 Convolutional neural network 、 Segmentation 、 Artificial intelligence
摘要: Accurate teeth segmentation in panoramic radiographic X-Ray images is importance for orthodontic treatment and research. This paper evaluates the method of LeNet-5 convolutional neural network with input data windowed image patches automated tooth root segmentation. In total, 103,984 created from 798 are used training validation sets. The proposed produced an accuracy 87.94%, which higher than comparative Sobel- Canny-processed cases. A visual evaluation shows a close resemblance to ground truth. achieved high performance on dental images. With some slight further modification improvement, might be applicable first step diagnosis or analysis systems, involves similar tasks.