A unified approach to salient object detection via low rank matrix recovery

By Xiaohui Shen, Ying Wu. 2012 IEEE Conference on Computer Vision and Pattern Recognition, 2012

719 citations recorded by the source corpus.

Subfields: Cognitive Neuroscience, Computer Vision and Pattern Recognition, Human-Computer Interaction

View Paper: https://www.semanticscholar.org/paper/db0cb86cfafb0405c46c45fd7d2d89bd5123740a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cognitive Neuroscience

1. Ming-Ming Cheng
2. Huchuan Lu
3. Ali Borji
4. Deng-Ping Fan
5. Lihe Zhang
6. Qingming Huang
7. Xiang Ruan
8. G. Wang

In Computer Vision and Pattern Recognition

1. Ming-Hsuan Yang
2. Ming-Ming Cheng
3. Huchuan Lu
4. Luc Van Gool
5. Dacheng Tao
6. G. Wang
7. Junwei Han
8. Ling Shao

In Human-Computer Interaction

1. Ming-Ming Cheng
2. Yizhou Yu
3. Guanbin Li
4. Haibin Ling
5. Yichen Wei
6. Junwei Han
7. Ali Borji
8. Jun Liu

Citations in the corpus

• Saliency Detection on Light Field (2017, positive): “To incorporate the location prior [20], we scale the focusness measure for each region Rr in terms of its distance to the center of the image and use it as a new background cue”
• Salient object detection: A survey (2014, background): “Based on such a general low-rank matrix recovery framework, Shen and Wu [177] propose a unified approach to incorporate traditional low-level features with higherlevel guidance, e.”
• Visual saliency based on multiscale deep features (2015, methodology): “A model based on low-rank matrix recovery is presented in [33] to integrate low-level visual features with higher-level priors.”
• Saliency Detection via Dense and Sparse Reconstruction (2013, methodology): “On the other hand, salient objects are likely to appear at the center of a scene [19, 5].”
• Deep Contrast Learning for Salient Object Detection (2016, background): “Salient object detection can be performed either in a bottom-up fashion using low-level features [14, 1, 30, 23, 39, 49, 21, 51, 8] or in a top-down fashion via the incorporation of high-level knowledge [22, 5, 17, 41, 29, 19, 28].”
• Detection of Co-salient Objects by Looking Deep and Wide (2016, methodology): “…et al. 2014), CBCS (Fu et al. 2013), ESMG (Li et al. 2015), CBCS-S 2 (Fu et al. 2013), ESMG-S 3 (Li et al. 2015), BLSM (Xie et al. 2013), and LR (Shen and Wu 2012), where the first six methods are proposed by state-of-the-art co-saliency detection methods and the last two methods are…”
• Saliency Detection via a Multiple Self-Weighted Graph-Based Manifold Ranking (2020, methodology): “Some methods simulate saliency detection in Bayesian inferencemodels [10], [11] and sparse reconstruction [12], [13].”
• Contextual Hypergraph Modelling for Salient Object Detection (2013, methodology): “These approaches model uniqueness by exploiting the global statistical properties of the image, including frequency spectrum analysis [4], color-spatial distribution modeling [7], high-dimensional Gaussian filtering [11], low-rank matrix decomposition [12], and geodesic distance computation [5].”
• Saliency Detection via Absorbing Markov Chain (2013, neutral): “MSRA: On the MSRA dataset, we compare the proposed method with eleven state-of-the-art methods which are LR [29], CB [17], SVO [7], RC [8], CA [27], SER [31], FT [2], GB [14], SR [15], LC [37] and IT [16].”
• Hierarchical Saliency Detection (2013, methodology): “The abbreviations are the same as those in [23], except for LR, which represents the low rank method of [25].”
• Minimum Barrier Salient Object Detection at 80 FPS (2015, methodology): “Secondly, similar to many previous methods [29, 12], to account for the center bias that is observed in many salient object detection datasets [3], we pixel-wise multiply S with a parameter-free centeredness map C = [c ]W×H , which is defined as”
• Visual Saliency Detection Based on Multiscale Deep CNN Features (2016, methodology): “A model based on low-rank matrix recovery is presented in [29] to integrate lowlevel visual features with higher-level priors.”
• Salient object detection via bootstrap learning (2015, methodology): “We first evaluate the proposed algorithms and other 19 state-of-the-art methods including the IT98 [16], SF [28], LRMR [30], wCO [43], GS SP [34], XL13 [36], RA10 [29], GB [12], LC [41], SR [15], FT [1], CA [11], SVO [8], CBsal [18], GMR [38], GC [10], HS [37], RC-J [9] and DSR [23] methods on the A”
• Hierarchical and Robust Convolutional Neural Network for Very High-Resolution Remote Sensing Object Detection (2019, background): “Traditional CNNs adopt only the highest level features, while Shen and Wu [50] have proven that combining features with different semantic level would be better.”
• Salient Region Detection by UFO: Uniqueness, Focusness and Objectness (2013, methodology): “In order to evaluate the performance on detecting multiobject saliency, we also compare our method with most recent ones [7, 29, 34, 37] on the BSD-300 dataset [25].”
• Deep networks for saliency detection via local estimation and global search (2015, neutral): “Saliency cues such as center and object bias [31, 22], contrast information [38] and background prior [33, 15] have been shown to be effective in previous work.”
• Hierarchical Image Saliency Detection on Extended CSSD (2016, methodology): “Our method yields improvement over others on the new extended CSSD dataset as well as other benchmark datasets.”