ICCV2013 录用论文（目标跟踪相关部分）（转）

单目标(表观模型)：

1. Seunghoon Hong, Bohyung
Han. Orderless Tracking
through Model-Averaged Density Estimation. （Offline tracking?和一般的object tracking还是不一样的。 CVPR12上也有篇Orderless
Tracking, 不过是online tracking）

2. Zhibin Hong, Xue Mei, Dacheng
Tao. Tracking via Robust Multi-Task Multi-View Joint Sparse Representation. （稀疏表示）

3. Naiyan Wang.Online
Robust Non-negative Dictionary Learning for Visual Tracking.

4. Jin Gao. Discriminant Tracking Using Tensor Representation with Semi-supervise d Improvement. （Tensor，中科院的强项来了）

5. Kwang Yi. Initialization-Insensitive Visual Tracking Through Voting with Salient Local Features.

6. Junliang Xing. Robust Object Tracking with Online Multi-lifespan Dictionary Learning.

7. Dapeng Che. Constructing Adaptive Complex Cells for Robust Visual Tracking.

8. Qinxun Bai.Randomized
Ensemble Tracking. （草草看了一下，作者在线学习了弱分类器的权重系数，目标用了分块提取的直方图特征。）

9. Stefan Duffner. PixelTrack: a fast adaptive algorithm for tracking non-rigid objects.

10. Martin Schiegg. Conservation Tracking.

多目标(关联跟踪)：

1. Caglayan Dicle. The Way They Move: Tracking Multiple Targets with Similar Appearance.

2. Baoyuan Wu, Ji
Qiang, and et al. Simultaneous Clustering and Tracklet Linking for Multi-Face Tracking in Videos.

3. Siyu Tang. Learning
People Detectors for Tracking in Crowded Scenes. (Tracking failure, occlusion pattern)

4. Chetan Arora. Higher Order Matching for Consistent Multiple Target Tracking.

5. Aleksandr Segal, Ian Reid. Latent
Data Association: Bayesian Model Selection for Multitarget Tracking.

跟踪算法评价：

1. YU PANG, Haibin Ling.Finding
the Best from the Second Bests -- Inhibiting Subjective Bias in Evaluation of Visual Tracking Algorithms. (这是一篇很有趣的文章，作者认为如果tracking paper做性能比较时一般都会有bias存在，因为调参、选视频和对比算法中不可避免加入作者主观的因素使proposed tracker在competition中胜出。然而，作者注意到了一点，除去那个proposed
tracker，其他trackers(也就是题目中的second bests)的比较结果的可信度应该比较大，所以可以搜集跟踪论文中除去proposed tracker的所有结果，从这些数据中李云排序算法得到各种跟踪算法的性能排名。最后结果Struck还是排名第一。)

从收录论文来看可见几种趋势，在RGB-D图像上的跟踪开始增多（上文没有列出），扫了一眼收录论文至少有3篇。基于Dictionary Learning的有2篇录用，从子空间（PCA）到sparse-coding-based tracking（L1）再到到Dictionary Learning-based tracking这个过程也相当的清晰了。至于多目标跟踪，需要多多学习了。

*以上对单/多目标的分类仅从论文题目上判断。

其他感兴趣的工作（陆续发现中）：

1. Michael Gygli, Helmut
Grabner.The Interestingness of Images.

2. Hamed Kiani galoogahi,Terence
Sim,Simon Lucey.Multi-Channel
Correlation Filters. (Minimum Output Sum of Squared Error Filter对于多通道特征的扩展)

3. Carl Vondrick, Aditya Khosla, Tomasz Malisiewicz, Antonio Torralba. HOGgles: Visualizing
Object Detection Features. (from MIT, demo可以把图像的HoG特征可视化，cool!)

4. Rui Zhao, Wanli Ouyang, Xiaogang
Wang. Person Re-identification by Salience Matching. (Re-Id, from CUHK)

5. Weilin Huang,  Zhe Lin,Jianchao
Yang,Jue Wang.Text
Localization in Natural Images using Stroke Feature Transform and Text Covariance Descriptors. (byAdobe<人家叫阿逗比，不是阿逗巴>, 笔画变换+Cov描述子)

6. Javier Marin, David Vazquez, Jaume Amores, Antonio Lopez, Bastian Leibe.Random
Forests of Local Experts for Pedestrian Detection. (随机森林, 行人检测)

7. Taegyu Lim, Seunghoon Hong, Bohyung Han, Joon Hee Han. Simultaneous Segmentation and Pose Tracking in Moving Camera.

8. Piotr Dollar, Larry Zitnick. Structured
Forests for Fast Edge Detection. （Random Forest用于边缘检测，新颖，快评见此）

9. Danhang Tang, T-K.
Kim. Real-time Articulated Hand Pose Estimation using Semi-supervised
Transductive Regression Forests.

10. X. Zhu, C. C. Loy, and S. Gong. Video
Synopsis by Heterogeneous Multi-Source Correlation. (视频摘要, Clustering Forest 应用)

11. Samuel Schulter, Christian Leistner, Paul Wohlhart and et al. Alternating
Regression Forests for Object Detection and Pose Estimation. (TU Graz)