R³Net: Recurrent Residual Refinement Network for Saliency Detection

By Zijun Deng, Xiaowei Hu, Lei Zhu, Xuemiao Xu, Jing Qin, Guoqiang Han, P. Heng. International Joint Conference on Artificial Intelligence, 2018

518 citations recorded by the source corpus.

Saliency detection is a fundamental yet challenging task in computer vision, aiming at highlighting the most visually distinctive objects in an image. We propose a novel recurrent residual refinement network (R^3Net) equipped with residual refinement blocks (RRBs) to more accurately detect salient regions of an input image. Our RRBs learn the residual between the intermediate saliency prediction and the ground truth by alternatively leveraging the low-level integrated features and the high-level integrated features of a fully convolutional network (FCN). While the low-level integrated features are capable of capturing more saliency details, the high-level integrated features can reduce non-salient regions in the intermediate prediction. Furthermore, the RRBs can obtain complementary saliency information of the intermediate prediction, and add the residual into the intermediate prediction to refine the saliency maps. We evaluate the proposed R^3Net on five widely-used saliency detection benchmarks by comparing it with 16 state-of-the-art saliency detectors. Experimental results show that our network outperforms our competitors in all the benchmark datasets.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/fcd6b89c55a636e042146a13711302cfe4aaf2bf

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 Media Technology

Citations in the corpus

Saliency Detection via a Multiple Self-Weighted Graph-Based Manifold Ranking (2020, background): “And then, multi-level convolutional features [33], [34] are aggregated for more accurately detect salient objects.”
Nested Network With Two-Stream Pyramid for Salient Object Detection in Optical Remote Sensing Images (2019, methodology): “We compare the proposed method with 14 state-of-the-art salient object detection methods on the testing subset of ORSSD data set, including 7 unsupervised methods (DSR [20], RBD [18], RRWR [48], HDCT [49], DSG [50], MILPS [51], and RCRR [15]), 4 deep learning-based methods (DSS [24], RADF [25], R3Ne”
Label Decoupling Framework for Salient Object Detection (2020, methodology): “To demonstrate the effectiveness of the proposed method, 14 state-of-the-art SOD methods are introduced to compare, including BMPM [38], DGRL [28], R(3)Net [7], RAS [4], PiCA-R [21], AFNet [13], BASNet [23], CPD-R [33], EGNet-R [41], PAGE [31], TDBU [30], SCRN [34], SIBA [24] and PoolNet [20].”
Suppress and Balance: A Simple Gated Network for Salient Object Detection (2020, methodology): “We adopt the atrous convolution with dilation rates of [2, 4, 6] and the same dilation rates are also applied to the folded convolution.”
F3Net: Fusion, Feedback and Focus for Salient Object Detection (2019, methodology): “(Deng et al. 2018) and (Wang et al. 2017b) adopted an iterative strategy to reﬁne the saliency maps step-by-step, using features both from deep layers and shallow layers.”
Weakly-Supervised Salient Object Detection via Scribble Annotations (2020, methodology): “Models Metric DGRL UCF PiCANet R3Net NLDF MSNet CPD AFNet PFAN PAGRN BASNet SBF WSI WSS MNL MSW Ours [35] [53] [21] [5] [23] [40] [41] [8] [56] [54] [26] [48] [16] [34] [52] [47]”
Target-aware Dual Adversarial Learning and a Multi-scenario Multi-Modality Benchmark to Fuse Infrared and Visible for Object Detection (2022, methodology): “We employ a pre-trained saliency detection network [3] to calculate the target mask m from infrared images so that the two discrim-inators can perform on their respective regions (target and background).”
BBS-Net: RGB-D Salient Object Detection with a Bifurcated Backbone Strategy Network (2020, methodology): “Note that our cascaded refinement mechanism is different from the recent refinement mechanisms R3Net [14], CRN [6], and RFCN [61] in its usage of multilevel features and the initial map.”
MovieNet: A Holistic Dataset for Movie Understanding (2020, methodology): “First, we adopt R(3)Net [17] to get the saliency map of each frame in the shot.”
Salient Object Detection via Integrity Learning (2021, methodology): “In Table 2, we compare our ICON with 16 state-of-the-art methods, including Amulet [22], DSS [35], NLDF [48], C2SNet [49], SRM [100], R3Net [101], BMPM [102], DGRL [103], PiCANet-R (PiCA-R) [25], RANet [104], AFNet [98], CPD [58], PoolNet [28], EGNet [51], BANet [50] and SCRN [55] in terms of…”
Global Context-Aware Progressive Aggregation Network for Salient Object Detection (2020, methodology): “For further enhance the saliency map, several recent works (Hou et al. 2017; Zhang et al. 2017; Deng et al. 2018; Hu et al. 2018; Li et al. 2018; Zhang et al. 2018a; 2018b) integrate features in multiple layers of CNN to exploit the context information at different semantic levels.”
Visual Saliency Transformer (2021, background): “Other works [36, 64, 11, 42, 6] tried to design recurrent networks to refine the saliency map step-by-step.”