中国科学院机构知识库网格系统: 检索

浏览/检索结果: 共6条，第1-6条

帮助

条数/页：排序方式：
	SimpleDet: A Simple and Versatile Distributed Framework for Object Detection and Instance Recognition 期刊论文 OAI收割 JOURNAL OF MACHINE LEARNING RESEARCH, 2019, 卷号: 20, 页码: 8 作者: Chen, Yuntao; Han, Chenxia; Li, Yanghao; Huang, Zehao; Jiang, Yi \| 收藏 \| 浏览/下载：60/0 \| 提交时间：2019/12/16 Object Detection Instance Recognition Distributed Training Mixed Precision Training
	Multi-Perspective Cost-Sensitive Context-Aware Multi-Instance Sparse Coding and Its Application to Sensitive Video Recognition 期刊论文 OAI收割 IEEE TRANSACTIONS ON MULTIMEDIA, 2016, 卷号: 18, 期号: 1, 页码: 76-89 作者: Hu, Weiming; Ding, Xinmiao; Li, Bing; Wang, Jianchao; Gao, Yan 收藏 \| 浏览/下载：57/0 \| 提交时间：2016/03/19 Cost-sensitive context-aware multi-instance sparse coding (MI-SC) horror video recognition multi-perspective multi-instance joint sparse coding (MI-J-SC) video emotional feature extraction violent video recognition
	Horror Image Recognition Based on Context-Aware Multi-Instance Learning 期刊论文 OAI收割 IEEE TRANSACTIONS ON IMAGE PROCESSING, 2015, 卷号: 24, 期号: 12, 页码: 5193-5205 作者: Li, Bing; Xiong, Weihua; Wu, Ou; Hu, Weiming; Maybank, Stephen 收藏 \| 浏览/下载：107/0 \| 提交时间：2015/11/12 Horror image recognition context-aware multi-instance learning visual saliency
	复杂场景中行为识别的关键技术及方法研究学位论文 OAI收割工学博士, 中国科学院自动化研究所: 中国科学院大学, 2014 作者: 周文收藏 \| 浏览/下载：105/0 \| 提交时间：2015/09/02 复杂场景视角局部特征多示例学习马尔科夫模型行为识别姿态估计 complex scenes viewpoints local features multiple-instance learning Markov model action recognition pose estimation
	Boosted Exemplar Learning for Action Recognition and Annotation 期刊论文 OAI收割 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, 2011, 卷号: 21, 期号: 7, 页码: 853-866 作者: Zhang, Tianzhu; Liu, Jing; Liu, Si; Xu, Changsheng; Lu, Hanqing 收藏 \| 浏览/下载：38/0 \| 提交时间：2015/08/12 Action annotation action recognition AdaBoost mi-SVM multiple instance learning (MIL)
	Adaptive resolution storage system based on LOG-POLAR transform for multi-target trackers (EI CONFERENCE) 会议论文 OAI收割 2010 International Conference on Computer Application and System Modeling, ICCASM 2010, October 22, 2010 - October 24, 2010, Shanxi, Taiyuan, China 作者: Zhang Y. 收藏 \| 浏览/下载：26/0 \| 提交时间：2013/03/25 The main constrained problem of the video monitoring and storage system is the contradiction between large field of view and storage space limitations. Not all of the video information introduced by the image sensors need to be recorded especially for some tracking system which has appointed functions to track specifically kinds of targets. For instance the system not only works for single target if the monitor is appointed for human face tracking but also can work for multi-targets. High reconstruction resolution in the fovea region enables the successive application of recognition modules without sacrificing their performance the best system appears to be concentrating on human face only and all the others considered being background the low reconstruction resolution in the periphery helps to reduce the video data. 2010 IEEE. the background regions needn't to be recorded in detail. For this purpose this letter presents a real-time foveate storage system which efficiently represents the video image in log-polar coordinates with the foveate point centered on the target