Learning to Detect A Salient Object

By Tie Liu, Zejian Yuan, Jingdong Wang, Nanning Zheng, Xiaoou Tang, Harry Shum. 2007

2,609 citations recorded by the source corpus.

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

How researchers see this paper

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

In Computer Vision and Pattern Recognition

In Human-Computer Interaction

Citations in the corpus

SSUL: Semantic Segmentation with Unknown Label for Exemplar-based Class-Incremental Learning (2021, methodology): “For the saliency-map detector to generate the Unknown class label, we employed DSS [11] pretrained on MSRA-B dataset [22], which contains 5,000 labels for salient objects.”
Rethinking the Smaller-Norm-Less-Informative Assumption in Channel Pruning of Convolution Layers (2018, methodology): “base model pruned model test dataset (#images) mIOU mIOU MSRA10K (Liu et al., 2011) (2,500) 83.”
Res2Net: A New Multi-Scale Backbone Architecture (2019, methodology): “Following [18], we train those two models using the MSRA-B dataset [27], and evaluate results on ECSSD [45], PASCAL-S [24], HKU-IS [23], and DUT-OMRON [46] datasets.”
Global contrast based salient region detection (2011, background): “[20] find multi-scale contrast by linearly combining contrast in a Gaussian image pyramid.”
Hierarchical Image Saliency Detection on Extended CSSD (2016, background): “Second, we build a new hierarchical inference model with a local consistency scheme, which leads to more natural saliency results compared to previous tree-structured model.”
Deep Contrast Learning for Salient Object Detection (2016, methodology): “While it takes around 25 hours to train our deep contrast network using the MSRA-B dataset, it only takes 1.5 seconds for the trained model (DCL) to detect salient objects in a testing image with 400x300 pixels on a PC with an NVIDIA Titan Black GPU and a 3.4GHz Intel processor.”
Deeply Supervised Salient Object Detection with Short Connections (2016, methodology): “We evaluate our approach on 5 representative datasets, including MSRA-B [43], ECSSD [54], HKU-IS [47], PAS-”
Saliency Detection via Dense and Sparse Reconstruction (2013, methodology): “The MSRA database [15] contains 5000 images, where each one is labeled with bounding boxes by nine subjects.”
Depth saliency based on anisotropic center-surround difference (2014, methodology): “Next, following the procedure in [20], we selected 5913 images each of which contains a salient object of moderate size.”
Semantic colorization with internet images (2011, methodology): “Salient foreground objects are segmented automatically from these images by applying the saliency detector in [Liu et al. 2007] and the Grabcut algorithm [Rother et al. 2004].”
Learning Pixel-Level Semantic Affinity with Image-Level Supervision for Weakly Supervised Semantic Segmentation (2018, background): “They substantially outperform the approaches based on the same level of supervision but with extra data and annotations like segmentation labels in MS-COCO [20], class-agnostic bounding boxes in MSRA Saliency [21], and YouTube videos [31].”
Deep Level Sets for Salient Object Detection (2017, methodology): “Different from bottom-up methods, top-down approaches [12, 34, 20] incorporate high-level visual knowledge into detection.”
Attentive Systems: A Survey (2017, methodology): “Salient foreground objects are segmented automatically from these images by applying the saliency detector in Liu et al. (2011) and the Grabcut algorithm (Rother et al. 2004).”
Exploiting Saliency for Object Segmentation from Image Level Labels (2017, methodology): “We generate segments from the MSRA boxes by applying grabcut over the average box annotation, and use these as supervision for the DeepLab model.”
Saliency Detection via Graph-Based Manifold Ranking (2013, background): “Bottom-up methods [1, 2, 6, 7, 9– 12, 14, 15, 17, 21, 24, 25, 27, 32, 33, 37] are data-driven and pre-attentive, while top-down methods [23, 36] are taskdriven that entails supervised learning with class labels.”
Cascaded Partial Decoder for Fast and Accurate Salient Object Detection (2019, background): “Besides, we also train the proposed model on MSRA-B dataset to compare with these three original models, and the results are reported in supplementary material.”
Measuring the Objectness of Image Windows (2011, background): “[36] detect a single dominant salient object in an image.”
Fast End-to-End Trainable Guided Filter (2018, methodology): “In Computer Vision, the pixelwise dense prediction techniques not only address the problem of segmenting an image into semantic parts [22, 35, 11], but also help to estimate depth from a single image [34], and detect the most salient object in an image [32, 29].”
Visual saliency detection by spatially weighted dissimilarity (2011, methodology): “Following the visual saliency model [4], the center-surround mechanism has been widely studied [7, 12].”
Frequency-tuned salient region detection (2009, methodology): “We derived the database from the publicly available database used by Liu et al. [20].”