Learning Uncertain Convolutional Features for Accurate Saliency Detection

By Pingping Zhang, Dong Wang, Huchuan Lu, Hongyu Wang, Baocai Yin. IEEE International Conference on Computer Vision, 2017

447 citations recorded by the source corpus.

Deep convolutional neural networks (CNNs) have delivered superior performance in many computer vision tasks. In this paper, we propose a novel deep fully convolutional network model for accurate salient object detection. The key contribution of this work is to learn deep uncertain convolutional features (UCF), which encourage the robustness and accuracy of saliency detection. We achieve this via introducing a reformulated dropout (R-dropout) after specific convolutional layers to construct an uncertain ensemble of internal feature units. In addition, we propose an effective hybrid upsampling method to reduce the checkerboard artifacts of deconvolution operators in our decoder network. The proposed methods can also be applied to other deep convolutional networks. Compared with existing saliency detection methods, the proposed UCF model is able to incorporate uncertainties for more accurate object boundary inference. Extensive experiments demonstrate that our proposed saliency model performs favorably against state-ofthe-art approaches. The uncertain feature learning mechanism as well as the upsampling method can significantly improve performance on other pixel-wise vision tasks.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/37f8e2749308294b31f0c222f703ef65af6e0e2b

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

Pyramid Dilated Deeper ConvLSTM for Video Salient Object Detection (2018, methodology): “We compared our model with 18 famous saliency methods, including 11 image salient object detection models: Amulet [51], SRM [36], UCF [52], DSS [16], MSR [23], NLDF [29], DCL [25], DHS [26], ELD [22], RFCN [35], KSR [37]; and”
Nested Network With Two-Stream Pyramid for Salient Object Detection in Optical Remote Sensing Images (2019, methodology): “[26] utilized an encoder fully convolutional network (FCN) and a corresponding decoder FCN to detect the salient object, in which a”
Salient Object Detection With Pyramid Attention and Salient Edges (2019, methodology): “Other methods leverage global and local saliency information [21, 34, 54, 29], combine pixel-and segment-level features [22], inspire connections between network layers [14], or explore more complex deep architectures [18, 26, 55, 37, 32].”
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]”
Review of Visual Saliency Detection With Comprehensive Information (2018, background): “The symmetrical network is also introduced in saliency detection, such as the encoder-decoder fully cnvolutional networks [44].”
Surface defect saliency of magnetic tile (2018, methodology): “Encoder-decoder structures generated from Fully Convolutional Network(FCN) [22] are mostly used in cuttingedge saliency detection models [17], [23], [24].”
U2-Net: Going Deeper with Nested U-Structure for Salient Object Detection (2020, background): “(UCF) [54] propose to reduce the checkerboard artifacts of deconvolution operators by introducing a reformulated dropout and a hybrid upsampling module.”
JL-DCF: Joint Learning and Densely-Cooperative Fusion Framework for RGB-D Salient Object Detection (2020, background): “Many SOD models have been developed under the assumption that the inputs are individual RGB/color images [21, 47, 66, 74–76] or sequences [56, 62, 67, 68].”
Detect Globally, Refine Locally: A Novel Approach to Saliency Detection (2018, methodology): “The increased performance benefits from the role that CWM plays in filtering out the noise and cluttered 3/27/2018 main _newlly.html file:///media/wtt/Experiment/desktop/vis/camera/main%20_newlly.html 1/1 ILSVRC2012_test_00000026 ILSVRC2012_test_00047403 ILSVRC2012_test_00057576 ILSVRC2012_test_0005”
A Simple Pooling-Based Design for Real-Time Salient Object Detection (2019, methodology): “UCF [45] NLDF [28] DSS [9] MSR [17] DHS [23] Size 224× 224 400× 300 400× 300 400× 300 224× 224”
Depth-Induced Multi-Scale Recurrent Attention Network for Saliency Detection (2019, methodology): “Although many works [10, 14, 22, 30, 44, 45, 47] have devoted to RGB saliency detection and have achieved appealing performance, they might fail when coping with complex scenarios, such as multiple or transparent objects, similar foreground and background, complex background and low-intensity enviro”
Reverse Attention for Salient Object Detection (2018, methodology): “We compare the proposed method with 10 state-of-the-art ones, including 9 recent CNN-based approaches, DCL+ [22], DHS [26], SSD [16], RFCN [39], DLS [10], NLDF [30], DSS and DSS+ [8], Amulet [45], UCF [46], and one conventional top approach, DRFI [13], where symbol “+” indicates that the network inc”
BASNet: Boundary-Aware Salient Object Detection (2019, methodology): “We compare our method with 15 state-of-the-art models, PiCANetR [39], BMPM [72], R3Net [6], PAGRN [76], RADF [19], DGRL [65], RAS [4], C2S [36], LFR [73], DSS [17], NLDF [41], SRM [64], Amulet [74], UCF [75], MDF [35].”
A Bi-Directional Message Passing Model for Salient Object Detection (2018, methodology): “Firstly, most previous FCN-based saliency detection models [29, 27, 38, 12] stack single-scale convolutional and max pooling layers sequentially to generate deep features.”
Interactive Two-Stream Decoder for Accurate and Fast Saliency Detection (2020, methodology): “We compare our model with 16 existing saliency detection models, including 10 nearly real-time models, such as RFCN [32], Amulet [45], UCF [46], NLDF [24], CKT [18], BMP [44], PCA [22], PAGE [35], DSS [12], EGNet [48] and 6 real-time models, such as RA [3], AFNet [8], BASNet [26], CPD [37], PoolNet ”
Pyramid Feature Attention Network for Saliency Detection (2019, methodology): “To be speciﬁc, (1) the proposed method not only highlights the correct salient object regions clearly, but also well suppresses the saliencies [42] 0.8508 0.0484 0.9249 0.0478 0.9200 0.0392 0.8806 0.0788 0.7740 0.0635 GRL[33] 0.8341 0.0509 0.9230 0.0446 0.9130 0.0377 0.8811 0.0799 0.7788 0.0632 PAGR”
EGNet: Edge Guidance Network for Salient Object Detection (2019, methodology): “In this section, we compare our proposed EGNet with 15 previous state-of-the-art methods, including DCL [28], DSS [17,18], NLDF [35], MSR [26], ELD [13], DHS [32], RFCN [48], UCF [62], Amulet [61], PAGR [63], Pi-CANet [33], SRM [49], DGRL [52], RAS [3] and C2S [29].”
R³Net: Recurrent Residual Refinement Network for Saliency Detection (2018, methodology): “The R3Net is trained on the MSRA10K dataset [Cheng et al., 2015], which is widely used for training the saliency models [Lee et al., 2016; Zhang et al., 2017a].”