Research on Intelligent Decision Method of Image Segmentation Based on Deep Learning Technology

Abstract

Image segmentation is the first step in image analysis and one of the most important links. Because image segmentation can process images into simpler and more characteristic forms for analysis. With the rapid development of deep learning technology and its wide application in the field of semantic segmentation, the effect of semantic segmentation has been significantly improved. Image semantic segmentation is one of the core tasks of computer vision, and its goal is to efficiently classify each pixel of an input image. In recent years, deep learning is the technology that has the most profound impact on the computer field. With the help of deep learning, image semantic segmentation tasks have achieved many results in the fields of autonomous driving, biomedicine, and augmented reality. Compared with image classification and object detection, semantic segmentation can provide richer image semantic information. However, there are many problems in semantic segmentation based on deep learning. Firstly, it is difficult to make semantic segmentation data sets, which has the problems of difficult training and high production cost; Secondly, the computation and network parameters of most algorithms are huge, which makes them unable to be applied to mobile devices with limited computing resources, which limits the development of semantic segmentation; Moreover, many algorithms do not make full use of the hardware resources of the computing platform to accelerate the running speed of the program. This paper analyzes and summarizes the image semantic segmentation methods based on deep neural network. According to different network training methods, the existing image semantic segmentation is divided into fully supervised learning image semantic segmentation and weakly supervised learning image semantic segmentation. The effects, advantages and disadvantages of representative algorithms in each method are compared and analyzed, it also expounds the contribution of deep neural network to the field of semantic segmentation.