基于深度学习的语义分割综述

Overview of semantic segmentation based on deep learning

语义分割是深度学习计算机视觉方面的核心领域,有着很深的研究价值。语义分割技术的发展在近几年趋于成熟,从传统的方法到基于卷积神经网络方法的突破,构建了端到端的语义分割深度学习神经网络算法。这些方法被用于人工智能当中,应用在无人驾驶,遥感影像检测,医疗影像研究等方面。基于对经典语义分割算法进行学习,每个经典算法都有自己的特点,值得在此一一总结阐述。文章将针对语义分割的发展,优秀算法的网络架构特色,应用的场景进行介绍,最后将对语义分割算法作小结和展望。

Semantic segmentation is the core field of deep learning computer vision,which has deep research value.The development of semantic segmentation technology has matured in recent years.From traditional methods to methods based on convolutional neural networks,an end-to-end deep learning neural network semantic segmentation algorithm is realized.These methods are used in artificial intelligence,applied in unmanned driving,remote sensing image detection,medical image research and so on.Based on the learning of classical semantic segmentation algorithms,each classical algorithm has its own characteristics,which is worthy of summarizing and elaborating here.This article will focus on the development of semantic segmentation,the network architecture features of excellent algorithms,and the application scenarios.Finally,it will summarize and look forward to semantic segmentation.