标题:Image object detection and semantic segmentation based on convolutional neural network
作者:Zhang L.; Sheng Z.; Li Y.; Sun Q.; Zhao Y.; Feng D.
作者机构:[Zhang, L] School of Mechanical and Automotive Engineering, Liaocheng University, Liaocheng, Shandong, China;[ Sheng, Z] School of Management, Wuhan D 更多
通讯作者:Sheng, Z(yushin@stu.cque.edu.cn)
通讯作者地址:[Sheng, Z] School of Management, Wuhan Donghu UniversityChina;
来源:Neural Computing and Applications
出版年:2019
DOI:10.1007/s00521-019-04491-4
关键词:AdaBoost; CNN; Image object detection; Semantic segmentation
摘要:In this paper, an unsupervised co-segmentation algorithm is proposed, which can be applied to the image with multiple foreground objects simultaneously and the background changes dramatically. The color edge image in RGB space is extracted for semantic extraction. This method can effectively distinguish foreground and background by recursively modeling the appearance distribution of pixels and regions. The coherence of image foreground and background model is enhanced by using the correlation between different image regions and image interior. Experimental results show that deep convolutional neural network can effectively realize semantic classification of scene images by end-to-end feature learning and achieve accurate semantic segmentation of scene images. © 2019, Springer-Verlag London Ltd., part of Springer Nature.
收录类别:SCOPUS
资源类型:期刊论文
原文链接:https://www.scopus.com/inward/record.uri?eid=2-s2.0-85073943238&doi=10.1007%2fs00521-019-04491-4&partnerID=40&md5=516ebae81be068d5a84ac886d96b9a73
TOP