Instance-Level Salient Object Segmentation

By Guanbin Li, Yuan Xie, Liang Lin, Yizhou Yu. Computer Vision and Pattern Recognition, 2017

295 citations recorded by the source corpus.

Image saliency detection has recently witnessed rapid progress due to deep convolutional neural networks. However, none of the existing methods is able to identify object instances in the detected salient regions. In this paper, we present a salient instance segmentation method that produces a saliency mask with distinct object instance labels for an input image. Our method consists of three steps, estimating saliency map, detecting salient object contours and identifying salient object instances. For the first two steps, we propose a multiscale saliency refinement network, which generates high-quality salient region masks and salient object contours. Once integrated with multiscale combinatorial grouping and a MAP-based subset optimization framework, our method can generate very promising salient object instance segmentation results. To promote further research and evaluation of salient instance segmentation, we also construct a new database of 1000 images and their pixelwise salient instance annotations. Experimental results demonstrate that our proposed method is capable of achieving state-of-the-art performance on all public benchmarks for salient region detection as well as on our new dataset for salient instance segmentation.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/4afc04da0d58df97fec5c0f912a9650ad5057b25

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

Citations in the corpus

A Bi-Directional Message Passing Model for Salient Object Detection (2018, background): “[14] put forward a top-down refinement network, where features from deeper layers are passed to the shallower ones.”
Structure-Measure: A New Way to Evaluate Foreground Maps (2017, background): “, RGB SOD [15,74,82,86], RGB-D SOD [23,26,73,84,85], RGBT SOD [80], light field SOD [36,60, 77], VSOD [33], 360 SOD, SID [41], Saliency Ranking [2], Co-SOD [22,24,79], and HR SOD [71]) community has begun to widely adopt this measure even in the camouflaged object detection [19,57,72,72] and medical”
Multi-Scale Interactive Network for Salient Object Detection (2020, methodology): “We compare the proposed algorithm with 23 state-of-the-art saliency detection methods, including the SRM [32], PiCANet [20], DGRL [34], R3Net [6], CapSal [50], BAS-Net [27], BANet [30], ICNet [36], CPD [41], Amulet [51], NLDF [23], MSRNet [15], UCF [52], DSS [13], PAGR[53], RAS [4], C2SNet [17], HRS”
Reverse Attention for Salient Object Detection (2018, methodology): “Attention model was designed to weight multi-scale features in [12,30].”
Rethinking RGB-D Salient Object Detection: Models, Data Sets, and Large-Scale Benchmarks (2019, methodology): “While existing SOD methods [19]–[35] have achieved remarkable success, most of them only rely on RGB images and ignore the important depth information, which is widely available in modern smartphones (e.”
Camouflaged Object Detection (2020, background): “…for (b) panoptic segmentation [30] (which detects generic objects [39,44] including stuff and things), (c) salient instance/object detection [16,33,61,76] (which detects objects that grasp human attention), and (d) the proposed camouﬂaged object detection task, where the goal is to detect…”
Attentive Feedback Network for Boundary-Aware Salient Object Detection (2019, methodology): “Similar to [14], we modify the VGG16 network into a fully convolutional network by casting away the last two fully-connected layers along with the last pooling layer.”
Concealed Object Detection (2021, background): “Given an input image (a), we present the ground-truth for (b) panoptic segmentation [5] (which detects generic objects [18], [19] including stuff and things), (c) instance level salient object detection [11], [20], and (d) the proposed concealed object detection task, where the goal is to detect obj”
Image Segmentation Using Deep Learning: A Survey (2020, methodology): “Other models use multiscale analysis for segmentation, among them Dynamic Multiscale Filters Network (DMNet) [56], Context Contrasted Network and gated multiscale aggregation (CCN) [57], Adaptive Pyramid Context Network (APC-Net) [58], MultiScale Context Intertwining (MSCI) [59], and salient object ”
Pyramid Dilated Deeper ConvLSTM for Video Salient Object Detection (2018, methodology): “First, we pre-train the spatial-learning part (including PDC module and base network) using two image saliency datasets: MSRA10K [6], and DUTOMRON [49], and one video dataset: the training set of DAVIS dataset [31].”
Salient Object Detection in the Deep Learning Era: An In-Depth Survey (2019, methodology): “Some MTL-based SOD methods train different tasks on the same architecture in tandem [55], [90]; some learn multi-domain knowledge simultaneously by incorporating different objective terms into the loss function [87], [91], [112]–[114]; while some utilize a branched network structure in which the…”
Salient object detection: A survey (2014, methodology): “The second approach, initially proposed in [159], is leveraging edge information”
Anabranch network for camouflaged object segmentation (2019, methodology): “More recent methods modiﬁed fully convolutional networks (FCNs) to compute a pixel-wise saliency map (Hou et al., 2017; Le and Sugimoto, 2017; Li et al., 2017a; Liu and Han, 2016; Wang et al., 2017b,a).”
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].”
A Simple Pooling-Based Design for Real-Time Salient Object Detection (2019, methodology): “By default, our ablation experiments are performed based on the VGG-16 backbone and the union set of MSRA-B [25] and HKU-IS [18] datasets as done in [17] unless special explanations.”
Salient Objects in Clutter: Bringing Salient Object Detection to the Foreground (2018, methodology): “Compared with the recent new (a) instance-level ILSO dataset [29] which is labeled with discontinue coarse boundaries, (c) MSCOCO dataset [24] which is labeled with polygons, our (b, d) SOC dataset is labeled with smooth fine boundaries.”