20160720-23专题: Applications to High-level Vision Tasks by Deep Learning

专题: Applications to High-level Vision Tasks by Deep Learning

报告嘉宾1：欧阳万里（香港中文大学）

报告时间：2016年7月20日（星期三）晚20:00（北京时间）

报告题目：物体检测中利用深度学习方法学习物体形变和更好特征的方法探索

主持人：李玺（浙江大学）

摘要：深度学习的有效性已经在机器视觉和模式识别各个领域。本次Webinar中，我们以物体检测为主要应用背景，介绍我们在深度学习方法上的探索。下面将介绍我们的几方面探索：1.如何学习更好的特征；2.如何学习物体部位的变化。

报告相关文献列表：

[1]. Wanli Ouyang, Xingyu Zeng, Xiaogang Wang, Shi Qiu, Ping Luo, Yonglong Tian, Hongsheng Li,Shuo Yang, Zhe Wang, Hongyang Li, Kun Wang, Junjie Yan, Chen-Change Loy, and Xiaoou Tang, DeepID-Net: Object Detection with Deformable Part Based Convolutional Neural Networks, TPAMI, accepted.

[2]. Wanli Ouyang, Xiaogang Wang,Cong Zhang, Xiaokang Yang. Factors in Finetuning Deep Model for Object Detection with Long-tail Distribution, CVPR 2016.

[3]. Wanli Ouyang, Hongyang Li, Xingyu Zeng, and Xiaogang Wang, "Learning Deep Representation with Large-scale Attributes", In Proc. ICCV 2015. 

讲者简介：

Wanli Ouyang received the PhD degree in the Department of Electronic Engineering, The Chinese University of Hong Kong, where he is now a research assistant professor. His research interests include image processing, computer vision and pattern recognition. ImageNet Large Scale Visual Recognition Challenge (ILSVRC) is one of the most important grand challenges in computer vision. The team led by him ranks No. 2 in the ILSVRC 2014 and 2015 for object detection from image and No. 1 in ILSVRC 2015 for object detection from video. He receives the best reviewer award of ICCV2015. He has been the reviewer of many top journals and conferences such as IEEE TPAMI, TIP, IJCV, TCAS, TSP, TITS, TMM, TNN, TKDE, CVPR, ICCV. He is a member of the IEEE.

报告嘉宾2：乔宇（中国科学院深圳先进技术研究院）

报告时间：2016年7月20日（星期三）晚20:30（北京时间）

报告题目：Deeply Understanding Human Action and Expression

主持人：王琦（西北工业大学）

摘要： Human action understanding is receiving extensive research interests in computer vision nowadays due to its wide applications in surveillance, human-computer interface, sports video analysis, and content based video retrieval. The challenges of action understanding come from background clutter, viewpoint changes, and motion and appearance variations. In this talk, I will report our continuous efforts (CVPR13, ICCV13, CVPR 14, ECCV 14, CVPR15, IJCV 15, CVPR16) to address these challenges in the last 4 years. These works range from mining middle level parts, multi-view encoding of local descriptors, hierarchical model, to utilizing deep models for action recognition and detection. Experimental results on large public datasets (e.g. UCF101, HMDB51) demonstrate the effectiveness of the proposed methods. In addition, I will give a brief overview on our group’s recent progresses on facial analysis and expression classification.

报告相关文献列表：

[1]. B. Zhang, L. Wang, Z. Wang, Yu Qiao, and H. Wang " Real-time Action Recognition with Enhanced Motion Vector CNNs," Proc. Int. Conf. Computer Vision and Pattern Recognition ( CVPR), 2016

[2]. L. Wang, Yu Qiao, X. Tang, and L. Van Gool " Actionness Estimation Using Hybrid Fully Convolutional Networks," Proc. Int. Conf. Computer Vision and Pattern Recognition ( CVPR), 2016 

[3]. Y. Wen, Z. Li, Yu Qiao, " Age-Invariant Deep Face Recogniton," Proc. Int. Conf. Computer Vision and Pattern Recognition ( CVPR), 2016 

[4]. W. Zhu, J. Hu, G. Sun, X. Cao, Yu Qiao "A Key Volume Mining Deep Framework for Action Recognition," Proc. Int. Conf. Computer Vision and Pattern Recognition ( CVPR), 2016

[5]. Kaipeng Zhang, Lianzhi Tan, Zhifeng Li, Yu Qiao "Gender and Smile Classification using Deep Convolutional Neural Networks," ChaLearn Looking at People (LAP) Workshop, CVPR, 2016. 

[6]. Limin Wang, Yu Qiao , Xiaoou Tang, "MoFAP: A Multi-level Representation for Action Recognition," International Journal of Computer Vision (IJCV), 2015 

[7]. L. Wang, Yu Qiao, X. Tang, " Action Recognition with Trajectory-Pooled Deep-Convolutional Descriptors," Proc. Int. Conf. Computer Vision and Pattern Recognition ( CVPR), 2015

讲者简介：

Yu Qiao is a professor with the Shenzhen Institutes of Advanced Technology (SIAT), the Chinese Academy of Science, and the deputy director of multimedia research lab. He has been a JSPS fellow and then a project assistant professor with the University of Tokyo from 2007 to 2010. His research interests include computer vision, speech processing, pattern recognition, and deep learning. He has published more than 110 papers in these fields. He received the Lu Jiaxi young researcher award from the Chinese Academy of Science.

报告嘉宾3：周博磊（MIT CSAIL实验室）

报告时间：2016年7月20日（星期三）晚21:00（北京时间）

报告题目：Understanding and Leveraging the Internal Representation of Convolutional Neural Networks.

主持人：王兴刚（华中科技大学）

报告摘要：Abstract: With the success of deep learning architectures such as convolutional neural networks (CNN) for visual processing and the access to image databases with millions of labeled examples (e.g., ImageNet, Places), the state of the art in computer vision is advancing rapidly. One important factor for continued progress is to understand the representations that are learned by the inner layers of these deep learning architectures. Here we have a comparison study to analyze the representations of the CNN trained on ImageNet for object recognition and the CNN trained on Places Database for scene recognition respectively. As scenes are composed of objects, the CNN for scene classification automatically learns to discover meaningful objects detectors. With object detectors emerging as a result of learning to recognize scenes, our work demonstrates that the same network can perform both scene recognition and object localization in a single forward-pass, without ever having been explicitly taught the notion and localization of the objects. The presentation is based on my work published in ICLR'15 and CVPR'16.

报告相关文献列表：

[1]. B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba Learning Deep Features for Discriminative Localization. Computer Vision and Pattern Recognition (CVPR), 2016

[2]. B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba Object Detectors Emerge in Deep Scene CNNs. International Conference on Learning Representations (ICLR), 2015

[3]. B. Zhou, A. Lapedriza, J. Xiao, A. Torralba, and A. Oliva. Learning Deep Features for Scene Recognition using Places Database. Advances in Neural Information Processing Systems 27 (NIPS), 2014.

讲者简介：

Biography: Bolei Zhou is the 4th-year PhD student in Computer Science and Artificial Intelligence Laboratory at MIT, advised by Prof. Antonio Torralba. His research interest is on computer vision and machine learning. His research interest is on deep learning and high-level vision & AI tasks such as scene understanding and visual recognition. He is the award recipient of Facebook Fellowship and Microsoft Research Asia Fellowship.