Mask R-CNN

By Kaiming He, Georgia Gkioxari, Piotr Doll'ar, Ross B. Girshick. 2017

Subfields: 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 Computer Vision and Pattern Recognition

Citations in the corpus

End-to-End Object Detection with Transformers (2020, methodology): “Similarly to the extension of Faster R-CNN [37] to Mask R-CNN [14], DETR can be naturally extended by adding a mask head on top of the decoder outputs.”
Do Better ImageNet Models Transfer Better? (2018, methodology): “[6, 61, 32, 30, 27]), they have never been systematically explored across network architectures.”
Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation (2020, methodology): “For example, Panoptic FPN [47] incorporates a semantic segmentation head to Mask R-CNN [31], while Porzi et al .”
Group Normalization (2018, methodology): “The restriction on batch sizes is more demanding in computer vision tasks including detection [12, 47, 18], segmentation [38, 18], video recognition [60, 6], and other highlevel systems built on them.”
UPSNet: A Unified Panoptic Segmentation Network (2019, methodology): “Most recently, Mask R-CNN [15] introduced a joint approach to both mask prediction and recognition where one of the two parallel heads are dedicated to each task.”
LVIS: A Dataset for Large Vocabulary Instance Segmentation (2019, methodology): “1, we test the performance of Mask RCNN [12] out-of-the-box, and show the importance of adjusting two inference-time hyper-parameters.”
Rescaling Egocentric Vision: Collection, Pipeline and Challenges for EPIC-KITCHENS-100 (2020, methodology): “The first is Mask R-CNN [38] trained on MSCOCO [4].”
Self-Training for Class-Incremental Semantic Segmentation (2020, background): “Mask R-CNN (He et al. 2017) obtains promising performance on instance segmentation for each object.”
CornerNet: Detecting Objects as Paired Keypoints (2018, background): “…are extensively used in one-stage detectors (Liu et al., 2016; Fu et al., 2017; Redmon and Farhadi, 2016; Lin et al., 2017), which can achieve results highly competitive with two-stage detectors (Ren et al., 2015; Girshick et al., 2014; Girshick, 2015; He et al., 2017) while being more eﬃcient.”
Deformable ConvNets V2: More Deformable, Better Results (2018, methodology): “On Mask R-CNN, 42.7% APbbox and 37.0% APmask scores are obtained with the enriched deformation modeling, which are respectively 2.4% and 2.0% higher than those of the DCNv1 baseline.”
Conditional Convolutions for Instance Segmentation (2020, methodology): “Following the common practice [14, 22, 37], our models are trained with split train2017 (115K images) and all the ablation experiments are evaluated on split val2017 (5K images).”
kPAM: KeyPoint Affordances for Category-Level Robotic Manipulation (2019, methodology): “The keypoint detection network [25] requires object instance segmentation as the input, and we integrate Mask R-CNN [8] into our manipulation pipeline to accomplish this step.”
Bottom-Up Object Detection by Grouping Extreme and Center Points (2019, background): “To put this result in a context, the state-of-the-art Mask RCNN [15] gets a mask AP of 37.”
SSAP: Single-Shot Instance Segmentation With Affinity Pyramid (2019, methodology): “n. Another work [2,32] combines the predictions of detection and semantic segmentation with a CRFasRNN [50] to generate instances. FICS [33] develops the position sensitive score map [13]. Mask R-CNN [20] extends Faster R-CNN [46] by adding a segmentation mask predicting branch on each Region of Int”
MLPerf Training Benchmark (2019, background): “Compared with Mask R-CNN (Huang et al., 2016) and other two-stage solutions, SSD trades speed for accuracy.”
Meta-Learning with Warped Gradient Descent (2019, background): “We show that increasing the capacity of the meta-learner by defining warp-layers as Residual Networks (He et al., 2017) improves performance on classification tasks (Section 4.”
Classifying, Segmenting, and Tracking Object Instances in Video with Mask Propagation (2019, methodology): “Model Classification cls head Mask R-CNN [16], ResNeXt-101 32x48d [30] cls head Localization bbox head Mask R-CNN [16] bbox head Segmentation mask head DeepLabv3 [9], Box2Seg [29] mask head Tracking tracking head UnOVOST [47], ReID Net [18, 31] mask propagation head Optical Flow PWC-Net [35] -”
Towards Bounding-Box Free Panoptic Segmentation (2020, methodology): “All these methods rely on Mask R-CNN [13] for predicting things.”
PIoU Loss: Towards Accurate Oriented Object Detection in Complex Environments (2020, methodology): “Other possible ways of extracting OBB include, fitting detected masks [3,10] and regressing OBB with anchor-free models [49], two new concepts in literature.”
Improving One-stage Visual Grounding by Recursive Sub-query Construction (2020, methodology): “In the ﬁrst stage, a number of region proposals are generated by an oﬀ-line module such as Edge-Box [53], selective search [39] or pretrained detectors [26,35,13].”