CVPR 2013 Recruitment Papers (target tracking section) _CVPR

Official link of the complete employment thesis: http://www.pamitc.org/cvpr13/program.php

This year CVPR has open access, it is really good for the public ah, especially for the benefit of my research little rookie

This year's paper on rgb-d camera applications and research is getting more and more.

Of course, they are still more concerned about the tracking aspects of the papers. From the author's point of view, most of them are Chinese, and there are quite a lot of famous cows in tracking, such as Ming-hsuan Yang,robert Collins,chunhua Shen. In addition, from the initial evaluation of the topic, Sparse representation heat in the decline, so haibin ling teachers do not have this article employed, and the pure tracking-by-detection almost disappeared.

The following is an excerpt from the tracking aspect of the recruitment paper:

Oral section:

Structure preserving Object tracking. Lu Zhang, Laurens van der Maaten

Tracking Sports players with context-conditioned Motion Models. Jingchen Liu, Peter Carr, Robert Collins, Yanxi Liu

Post section:

Online Object tracking:a Benchmark. Yi Wu, Jongwoo Lim,ming-hsuan Yang (target tracking test video is really getting out of the real world, why not more from some actual application scenarios.) In addition, still extremely admire the author's patience and contribution, he/she < light from the name really bad Judge Ah > used two kinds of metric to comprehensively evaluate state-of-the-art trackers. It uses the AUC curve < records under the different overlap rate or the center localization error limit tracking sucess ratio> I think this is a very rigorous approach, which is probably the first one. Moreover, different initialization < simulation of the position error of the output of the detector, and the performance of the trackers in the different stages of the video to track the Test > were performed, which was hardly ever done before. Do this kind of test still need to invest a lot of time, just let a variety of codes on their own machine run up sometimes will let a person very crazy, and the tedious also only have done people can really experience it. Spoilers, the final result No.1 belong to struck)

Learning Compact Binary codes for Visual tracking. Xi Li,chunhua Shen, Anthony Dick, Anton van den Hengel (continuation of the author's previous thesis system, including using SVM to trace and add graph structure to describe the relationship between the samples. While putting binary codes in the title, the author did not clearly explain why the two-valued feature was used. This allows beginners like me to understand the cost of such a great effort to convert real value features to binary strings of the proceeds. ）

PART-BASED Visual tracking with Online latent structural Learning. Rui Yao, Qinfeng Shi,chunhua Shen, yanning Zhang, Anton van den Hengel

Self-Paced learning for long-term tracking. James Supancic III, Deva Ramanan (long-term's gimmick is still fascinating, like the tld of the year, to see if there are more ideas for engineering. After reading the feeling is the academic taste is heavier, although and TLD is the same highlight the long-term, but the author uses the increment SVM to train the detector, and uses the idea which learns from the key frame. For the essay, please go to Bfcat's blog. My feeling is that 2006-2010 is based on the boosting tracking of the golden Age, the essence is based on the online feature selection of the discriminant view. Then, SVM began to rise, of course, there are SVM tracking, may be due to SVM for small samples of high dimensional characteristics of the classification advantage. Representative paper can refer to: struck, it in the above the online Object tracking:a benchmark a number of evaluation to obtain the first, and provide the source code (c + + style, concise, very good). ）

Visual tracking via locality sensitive Histograms.shengfeng He, Qingxiong Yang, Rynson Lau, Jiang Wang, Ming-hsuan Yang (C Ityu of HK, using histograms as a table view in the current research background is really the opposite.

Minimum uncertainty Gap for robust Visual tracking.junseok Kwon, Kyoung Mu Lee (vtd author, browse the paper, full screen formula derivation Ah, no momentum to look down)

least soft-thresold squares tracking. Dong Wang,huchuan Lu,ming-hsuan Yang (originally on this article is very expected, but after watching the feelings of deception ... And the author's online object tracking with Sparse prototypes are almost identical in terms of the objective function and the optimization method, but the narrative angle is different. Not clear)

Tracking people and their Objects. Tobias Baumgartner,dennis Mitzel, Bastian leibe (This should also have application background and foreground)

(Not all of the above include multiple target tracking papers)

Other Articles of interest:

Alternating Decision forests.samuel schulter, Paul Wohlhart,christian leistner, Amir saffari,peter M. Roth,horst Bischof ( Forest is also one of the hot spots in recent years. In addition, the TU Graz Visual Group is active at various top meetings. The author introduces the idea of boosting into the RF, and the difficulty degree of sample classification as the sample weight.

semi-supervised Node splitting for Random Forest Construction.xiao Liu, Mingli Song, Dacheng Tao,zicheng Liu, luming Zhang , Chun Chen, Jiajun Bu (although forest is very hot, but there are few studies on forest at such meetings.) Because the RF in each node to split the need for each category of the distribution has a relatively stable statistics, if the labeled sample less, then this statistic is not stable, leading to find the splitting parameters are not good. Therefore, the author uses labeled+unlabeled data to solve this problem, using KDE to estimate the probability that the unlabeled sample belongs to a certain category. ）

Optimizing 1-nearest Prototype classifiers.paul Wohlhart, Michael donoser, Peter roth,horst Bischof (by Tu Graz) (I admire Tu Gra Z This group of Researchers,1-nn can play a flower, the thesis uses the optimized method to find the prototypes in a group of samples, so that each NN test only need to calculate distance with prototypes, The author adds that this method can naturally join metric learning.)

Single-pedestrian detection aided by Multi-pedestrian detection. Wanli Ouyang, Xiaogang Wang

Measuring crowd collectiveness.bolei zhou,xiaogang Wang (the top group in group behavioral Analysis, a calf)

Crossing the Line:crowd counting by-Integer programming with local Features. Zheng Ma,antoni Chan

Histograms of Sparse codes for Object Detection.xiaofeng Ren,deva Ramanan (feature learning, can provide richer texture descriptions than hog)

Multi-Source multi-scale counting in extremely dense crowd. Haroon Idrees, Imran Saleemi, Mubarak Shah

Human Pose estimation from still Images using body Parts Dependent joint Regressors.matthias dantone,juergen Gall, Luc Van Gool, Christian Leistner (Gall Hough is quite impressive)

Efficient detector adaptation for Object detection in a video.pramod Sharma, Ram Nevatia (with online training adaptive classifier to improve BAS Eline classifier, in which positive and negative samples of the online collection are an important point. In a word, is to use baseline classifier (low threshold) first detect the target, and then use adaptive classifier (paper is ferns) to determine whether it is false alarm

Robust multi-resolution pedestrian detection in traffic scenes. Junjie Yan, Xucong Zhang, Zhen Lei, Dong Yi, Shengcai Liao, Stan Li (Oral)

Fast Multiple-part based object detection using Kd-ferns. Dan Levi, Shai Silberstein, Aharon Bar-hillel

Online dominant and anomalous Behavior detection in videos. Mehrsan Javan Roshtkhari, Martin Levine

BoF meets Hog:feature extraction based on histograms of-oriented P.D.F for Image gradients. Takumi Kobayashi

The Svm-minus similarity Score for video face Recognition.lior Wolf, Noga Levy

Occlusion patterns for object class detection. Bojan Pepikj, Michael Stark, Peter gehler, Bernt Schiele

Separable Dictionary Learning. Simon Hawe, Matthias Seibert, Martin kleinsteuber

Scalable Sparse subspace Clustering. Xi Peng, Lei Zhang, Zhang Yi

from:http://blog.csdn.net/gxf1027/article/details/8650878

http://blog.csdn.net/loadstar_kun/article/details/22717357