Salient Object Detection: A Discriminative Regional Feature Integration Approach

By Huaizu Jiang, Zejian Yuan, Ming-Ming Cheng, Yihong Gong, Nanning Zheng, Jingdong Wang. 2013

1,301 citations recorded by the source corpus.

Feature integration provides a computational framework for saliency detection, and a lot of hand-crafted integration rules have been developed. In this paper, we present a principled extension, supervised feature integration, which learns a random forest regressor to discriminatively integrate the saliency features for saliency computation. In addition to contrast features, we introduce regional object-sensitive descriptors: the objectness descriptor characterizing the common spatial and appearance property of the salient object, and the image-specific backgroundness descriptor characterizing the appearance of the background of a specific image, which are shown more important for estimating the saliency. To the best of our knowledge, our supervised feature integration framework is the first successful approach to perform the integration over the saliency features for salient object detection, and outperforms the integration approach over the saliency maps. Together with fusing the multi-level regional saliency maps to impose the spatial saliency consistency, our approach significantly outperforms state-of-the-art methods on seven benchmark datasets. We also discuss several followup works which jointly learn the representation and the saliency map using deep learning.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

SSUL: Semantic Segmentation with Unknown Label for Exemplar-based Class-Incremental Learning (2021, methodology): “Saliency-map detector Weight transfer Memory sampling Methods 0-15 16-20 all Methods 0-15 16-20 all Methods 0-15 16-20 all DRFI [13] 76.”
Res2Net: A New Multi-Scale Backbone Architecture (2019, background): “Early approaches [2] utilize handcrafted representations of global contrast [7] or multi-scale region features [41].”
A Bi-Directional Message Passing Model for Salient Object Detection (2018, methodology): “We compare the proposed saliency detection model with 13 state-of-the-art methods, including 11 deep learning based methods (LEGS [27], MDF [15], RFCN [29], ELD [12], DCL [16], DHS [18], NLDF [20], DSS [8], Amulet [37], UCF [38], SRM [30]) and 2 conventional ones (DRFI [9] and BSCA [24]).”
Deep Stacked Hierarchical Multi-Patch Network for Image Deblurring (2019, neutral): “Model [8] [30] [15] [7] [31] OURS Fβ .”
EGNet: Edge Guidance Network for Salient Object Detection (2019, methodology): “We have evaluated the proposed architecture on six widely used public benchmark datasets: ECSSD [56], PASCAL-S [30], DUT-OMRON [57], SOD [36, 44], HKUIS [27], DUTS [46].”
A deep multi-level network for saliency prediction (2016, background): “Others have concentrated on highlighting the most important object regions in an image [3], [4], [5].”
Saliency Detection on Light Field (2017, neutral): “Similar foreground and background or complex background imposes challenges on state-of-the-art algorithms (e.g., RC [6], DRFI [7]).”
Single-Stage Semantic Segmentation From Image Labels (2020, background): “[54] iteratively train a network with the seed-and-expand principle, refine the intermediate results with saliency maps [53], and provide them as supervision to the segmentation network.”
Minimum Barrier Salient Object Detection at 80 FPS (2015, methodology): “We provide an extension of the proposed method by integrating the appearance-based backgroundness cue [16], which assumes that background regions are likely to possess similar appearance to the image boundary regions.”
New Contour Cue-Based Hybrid Sparse Learning for Salient Object Detection (2019, methodology): “Jiang et al. [38] regarded saliency map generation process as a regression task.”
Self-Erasing Network for Integral Object Attention (2018, methodology): “Saliency cues [5, 2, 33, 10, 13] are also adopted by some methods as the initial heuristic cues.”
Fast End-to-End Trainable Guided Filter (2018, methodology): “KITTI [17], MSRA-B [26] and PASCAL VOC 2012 [12] are used for training and testing respectively.”
Surface defect saliency of magnetic tile (2018, methodology): “GMR [28], MBP [29], MC [30], RBD [31] and DRFI [32] are top models of non-deeplearning models in natural scene datasets [27] , so they are used to compare to our MCue model.”
Recurrent Attentional Networks for Saliency Detection (2016, methodology): “For the HKUIS and ECSSD dataset, the F-measure improvements of the proposed method over the next top-performing method DRFI are more than 5%.”
Deep Level Sets for Salient Object Detection (2017, methodology): “Comparison between the proposed method and methods FT [1], HC [10], DRFI [21], GMR [56], BSCA [47],MCDL [62], LEGS [50], MTDS [31], MDF [29], ELD [26].”
Saliency detection by multi-context deep learning (2015, methodology): “A large number of approaches [63, 52, 40, 39, 32, 35, 60, 57, 56, 47, 41, 31, 27, 25, 24, 23, 11, 44, 17, 8, 13, 1, 21] are proposed to capture different saliency cues.”
Learning Integral Objects With Intra-Class Discriminator for Weakly-Supervised Semantic Segmentation (2020, background): “Meanwhile, external saliency models [17, 13] are required to estimate the background, which implicitly introduces additional pixel-level annotation requirements.”
Weakly-Supervised Semantic Segmentation Network with Deep Seeded Region Growing (2018, methodology): “For localizing background, we utilize the saliency detection technology from [12], and simply”