ICCV 2013论文出炉喽
ICCV 13论文：http://www.cs.toronto.edu/~kyros/local_outgoing/ICCV-Final-Results/
单是标题含有saliency/salient的论文就有14篇，做显著性检测的小伙伴们加油哦
ICCV saliency检测相关的论文列表：
1. (oral)Benchmarking Computational Model of Visual Saliency. Ali Borji*, ; Dicky Sihite, University of Southern California (USC); Hamed Rezazadegan Tavakoli, University of Oulu; Laurent Itti, University of Southern California (USC);
2. (oral)Estimating Human Scanpath Using Hidden Markov Model（赶脚很相关）,Huiying Liu*, Institute of Computing Technol; Dong Xu, "NTU, Singapore"; Qingming Huang, Graduate Univ of Chinese Academy of Sciences; Wen LI, NTU; Stephen Lin, Microsoft Research Asia;
3. (poster)Image Aided Automatic Registration of Laser Scans via Salient Directions. Bernhard Zeisl*, ETH; Kevin Koeser, ; Marc Pollefeys, ETH
4. (poster)Efficient Salient Region Detection with Soft Image Abstraction. Ming-Ming Cheng*, Oxford Brookes University; Shuai Zheng, Oxford Brookes University; Jonathan Warrell, Oxford Brookes University; Vibhav Vineet, Oxford Brookes University; Wenyan Lin, Oxford Brookes University
5. (poster)Saliency Detection: A Boolean Map Approach. Jianming Zhang*, Boston University; Stan Sclaroff, Boston University
6. (poster)Salient Region Detection by UFO: Uniqueness, Focusness and Objectness. Peng Jiang*, Shandong University; Jingliang Peng, cs.sdu.edu.cn; Haibin Ling,
通常，对象的sharp edges投影到成像平面后，会变得模糊。聚焦和散焦对对象边缘的影响要大于对象内部区域，因而，可以通过获取区域的边界的focusness来反映区域显著性。其余两个不解释。同一篇论文中还用了两种over segmentation方法呢。另http://www.shawnlankton.com/2007/12/3d-vision-with-stereo-disparity/对run demo有用。另外，还需要将opencv_core220.dll, opencv_highgui220.dll, opencv_imgproc220.dll丢到根目录下，作者给出的源程序才能run。
7. (poster)Contextual Hypergraph Modeling for Salient Object Detection. Xi Li*, University of Adelaide; Yao Li, University of Adelaide; Chunhua Shen, The University of Adelaide; Anthony Dick, University of Adelaide ; Anton Van den Hengel, University of Adelaide
8. (poster)Initialization-Insensitive Visual Tracking Through Voting with Salient Local Features. Kwang Yi*, Seoul National University; Hawook Jeong, Seoul National University; Byeongho Heo, Seoul National University; Hyung Jin Chang, Imperial College London; Jin Young Choi, Seoul National University
9. (poster)Saliency and Human Fixations: State-of-the-art and Study of Comparison Metrics. Nicolas Riche*, UMONS; Matthieu Duvinage, UMONS; Matei Mancas, UMONS; Bernard Gosselin, UMONS; Thierry Dutoit, UMONS
10. (poster)Person Re-identification by Salience Matching. Rui Zhao*, CUHK; Wanli Ouyang, The Chinese University of HK; Xiaogang Wang, "The Chinese University of Hong Kong, Hongkong"
11. (poster)Saliency Detection via Dense and Sparse Reconstruction. Xiaohui Li*, DUT, China; huchuan Lu, DUT,China; Ming-Hsuan Yang, "UC Merced, USA"; Lihe Zhang, DUT, China; Xiang Ruan
作者利用位于图像四个边界的超像素做成背景模板。dense reconstruction就是对背景模板进行PCA，得到基向量，对图像中其余超像素，用基向量进行重构，重构误差小，说明超像素越像背景，重构误差大，说明超像素越不像背景。sparse reconstruction就是将背景模板做为字典原子，将图像中其余超像素用字典原子进行稀疏表示，同样可以得到重构误差。对两类重构误差，分别进行context-based error propogation，并且还是用了多尺度，对象偏向性修正，最后采用贝叶斯将显著图进行整合。这里的贝叶斯整合是将dense reconstruction得到的显著图做为先验，将sparse reconstruction得到的显著图做为似然，在计算后验概率。反过来可以得到另一个后验概率。两个后验概率相加得到最终的显著值。
12. (poster)Saliency Detection via Absorbing Markov Chain. Bowen Jiang*, DUT; Lihe Zhang, DUT, China; huchuan Lu, DUT,China; Ming-Hsuan Yang, "UC Merced, USA"; Chuan Yang,
对saliency object detection 问题的formulation很有意思，将absorbing Markov chain很好地用于salient object detection，立意新奇，prof. lu的一系列论文中model虽然简单，却是小清新范儿十足。本文方法与geodesic saliency一样，十分受边界object困扰。
13. (poster)Category-Independent Object-level Saliency Detection. Yangqing Jia*, UC Berkeley; Mei Han, Google Research
14. Saliency Detection in Large Point Sets. Elizabeth Shtrom*, Technion; George Leifman, Technion; Ayellet Tal, Technion
2013是不平凡的一年。
这一年，ASD达到史前最辉煌；
此后，ASD时代结束，MSRA5000时代来临。