中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 模式分析与学习团队
LG-CNN: From local parts to global discrimination for fine-grained recognition

文献类型：期刊论文

作者Xie, Guo-Sen1; Zhang, Xu-Yao1; Yang, Wenhan3; Xu, Mingliang4; Yan, Shuicheng5; Liu, Cheng-Lin1,2
刊名PATTERN RECOGNITION
出版日期2017-11-01
卷号71期号:71页码:118-131
关键词Fine-grained Recognition Convolutional Neural Networks Bilinear Pooling Local Parts Global Discrimination
DOI10.1016/j.patcog.2017.06.002
文献子类Article
英文摘要Fine-grained recognition is one of the most difficult topics in visual recognition, which aims at distinguishing confusing categories such as bird species within a genus. The information of part and bounding boxes in fine-grained images is very important for improving the performance. However, in real applications, the part and/or bounding box annotations may not exist. This makes fine-grained recognition a challenging problem. In this paper, we propose a jointly trained Convolutional Neural Network (CNN) architecture to solve the fine-grained recognition problem without using part and bounding box information. In this framework, we first detect part candidates by calculating the gradients of feature maps of a trained CNN model w.r.t the input image and then filter out unnecessary ones by fusing two saliency detection methods. Meanwhile, two groups of global object locations are obtained based on the saliency detection methods and a segmentation method. With the filtered part candidates and approximate object locations as inputs, we construct the CNN architecture with local parts and global discrimination (LG-CNN) which consists of two CNN networks with shared weights. The upper stream of LG-CNN is focused on the part information of the input image, the bottom stream of LG-CNN is focused on the global input image. LG-CNN is jointly trained by two stream loss functions to guide the updating of the shared weights. Experiments on three popular fine-grained datasets well validate the effectiveness of our proposed LG-CNN architecture. Applying our LG-CNN architecture to generic object recognition datasets also yields superior performance over the directly fine-tuned CNN architecture with a large margin. (C) 2017 Elsevier Ltd. All rights reserved.
WOS关键词IMAGE CLASSIFICATION ; FISHER VECTOR ; NETWORKS ; FEATURES
WOS研究方向Computer Science ; Engineering
语种英语
WOS记录号WOS:000406987400010
资助机构National Basic Research Program of China (973 Program)(2012CB316302) ; Strategic Priority Research Program of the CAS(XDA06040102 ; Natural Science Foundation of China(NSFC)(61472370 ; open project of State Key Laboratory of Virtual Reality Technology and System(BUAA-VR-16KF-07) ; XDB02060009) ; 61672469)
源URL[http://ir.ia.ac.cn/handle/173211/20356]  
专题自动化研究所_模式识别国家重点实验室_模式分析与学习团队
作者单位1.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 100190, Peoples R China
2.Univ Chinese Acad Sci, Beijing, Peoples R China
3.Peking Univ, Beijing 100871, Peoples R China
4.Zhengzhou Univ, Ctr Interdisciplinary Informat Sci Res, Zhengzhou 450001, Henan, Peoples R China
5.Natl Univ Singapore, Singapore 119077, Singapore
推荐引用方式
GB/T 7714		 Xie, Guo-Sen,Zhang, Xu-Yao,Yang, Wenhan,et al. LG-CNN: From local parts to global discrimination for fine-grained recognition[J]. PATTERN RECOGNITION,2017,71(71):118-131.
APA Xie, Guo-Sen,Zhang, Xu-Yao,Yang, Wenhan,Xu, Mingliang,Yan, Shuicheng,&Liu, Cheng-Lin.(2017).LG-CNN: From local parts to global discrimination for fine-grained recognition.PATTERN RECOGNITION,71(71),118-131.
MLA Xie, Guo-Sen,et al."LG-CNN: From local parts to global discrimination for fine-grained recognition".PATTERN RECOGNITION 71.71(2017):118-131.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。