基于卷积神经网络的目标检测算法综述

Survey of Object Detection Algorithms Based on Convolutional Neural Networks

目标检测作为计算机视觉中最基本和最具挑战性的核心任务之一,在国际上一直处于热门研究领域。近年来,随着深度学习的快速发展,基于卷积神经网络的目标检测算法几乎已经完全代替了传统目标检测算法。在简单的介绍了传统目标检测算法并且分析了其不足的基础上,介绍了卷积神经网络的发展,重点介绍了基于卷积神经网络的一阶段和两阶段目标检测算法,同时对各算法在VOC2012和COCO数据集上的性能表现进行总结。最后提出了目前目标检测存在的问题和解决办法,对目标检测未来发展方向进行了展望。

As one of the most basic and most challenging core tasks in computer vision,object detection has always been a hot research field in the world.In recent years,with the rapid development of deep learning,object detection algorithms based on convolutional neural networks have almost completely replaced traditional object detection algorithms.On the basis of a brief introduction of traditional object detection algorithms and analysis of their shortcomings,the development of convolutional neural networks is introduced,and one-stage and two-stage object detection algorithms based on convolutional neural networks are introduced.At the same time,the performance of each algorithm on the VOC2012 and COCO data sets is summarized.Finally,the current problems and solutions of object detection are proposed,and the future development direction of object detection is prospected.