摘要
由于光照变化、相机抖动和动态背景等因素影响,现有基于传统图像处理方法的前景提取算法并不能在复杂场景下获得良好的分割效果。针对此类问题,本文提出了一种基于全卷积孪生神经网络的前景提取算法,仅需任意2帧图像即可准确提取运动前景。将输入的2帧图像分为背景图像与待提取图像,将其输入全卷积孪生神经网络得到二者的相似性度量图,该相似性度量图中包含待提取图像相对于背景图像的各像素变化情况信息;接着将相似性度量图与待提取图像融合,利用编解码网络以实现端到端的前景提取。在CDnet2014数据集上进行综合评估与测试,结果均证明了该方法的有效性。
Due to factors such as illumination changes,camera jitter and dynamic background,existing foreground subtraction algorithms cannot achieve good segmentation results in complex scenes.To solve this kind of problems,this paper proposes a subtraction algorithm based on a fully convolutional siamese neural network,which can accurately segment the foreground with only two arbitrary frames.Specifically,the input two images are divided into the base image and the image to be segmented.The algorithm uses the fullyconvolutional siamese network to get the similarity metric map of input frames.The similarity metric map contains information about changes in pixels of the image to be segmented relative to the base image.Then,the similarity metric map is fused with the image to be segmented,and the encoder-decoder network is used to achieve end-to-end foreground subtraction results.The paper evaluates the proposed algorithm on the CDnet2014 dataset to prove its effectiveness.
作者
刘峰
居昊
干宗良
LIU Feng;JU Hao;GAN Zongliang(Jiangsu Key Laboratory of Image Processing and Image Communication,Nanjing 210003,China;College of Telecommunications and Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;College of Education Science and Technology,Nanjing University of Posts and Telecommunications,Nanjing 210003,China)
出处
《南京航空航天大学学报》
CAS
CSCD
北大核心
2021年第5期743-750,共8页
Journal of Nanjing University of Aeronautics & Astronautics
关键词
前景提取
度量学习
孪生神经网络
复杂监控场景
foreground subtraction
metric learning
siamese network
complex surveillance videos
