Efficient hierarchical graph-based video segmentation

By Matthias Grundmann, Vivek Kwatra, Mei Han, Irfan Essa. 2010

775 citations recorded by the source corpus.

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

Open-World Instance Segmentation: Exploiting Pseudo Ground Truth From Learned Pairwise Affinity (2022, methodology): “, [1, 4, 6, 18, 51, 58]), before deep learning for supervised object detection took off, were quite category-agnostic.”
Unsupervised Learning of Object Keypoints for Perception and Control (2019, methodology): “Both instance and semantic segmentation has been approached using supervised [22, 25] and unsupervised learning [2, 9, 14, 10, 16, 21, 6] methods.”
Unidentified Video Objects: A Benchmark for Dense, Open-World Segmentation (2021, methodology): “On YTVIS, GBH achieves 34.7% for approximate AR 100 , 5.9% higher than open-world YTVIS (train on VOC classes and test on all) and 6.3% lower than closed-world YTVIS (Table 6).”
One-Shot Video Object Segmentation (2016, methodology): “In the latter to JOTS [52], TMFG [25] KEY [24], HVS [17] HBT [16], ST [51] and OFL [49].”
Learning Video Object Segmentation from Static Images (2016, background): “Unsupervised segmentation Another family of work does general moving object segmentation (over all parts of the image), and selects post-hoc the space-time tube that best match the annotation, e.g. NLC [14] and [16,24,48].”
On Face Segmentation, Face Swapping, and Face Perception (2017, methodology): “Given a video, we produce a rough, initial segmentation using a method based on [10].”
Video Segmentation by Non-Local Consensus voting (2014, methodology): “Advancement in handling non-rigid motions has been made by supervoxel-based methods [3, 10, 15, 21, 36].”
Multi-class cosegmentation (2012, methodology): “This is confirmed with our experiments with two short video clips taken from the Hollywood-2 and Grundmann datasets [24, 25].”
Growth Signatures of Rosette Plants from Time-Lapse Video (2015, background): “builds a region graph and computes a tree of spatiotemporal segmentations [47].”
Superpixels: An evaluation of the state-of-the-art (2016, methodology): “Finally, several superpixel algorithms (among others [29], [30] and [31]) have been adapted to videos and image volumes – a survey and comparison of some of these so-called supervoxel algorithms can be found in [32].”
Advances in Computer Vision-Based Civil Infrastructure Inspection and Monitoring (2019, background): “Optical flow has wide ranging applications in processing video data, from video compression [53], to video segmentation [54], motion magnification [55] and vision-based navigation of UAVs [56].”
Stixmantics: A Medium-Level Model for Real-Time Semantic Scene Understanding (2014, background): “[24], have the potential to yield best possible segmentation performance, as they can access all available information from all time steps during inference.”
Video Object Segmentation with Joint Re-identification and Attention-Aware Mask Propagation (2018, background): “Prior to the prevalence of deep learning, most approaches to semantic video segmentation are graph based [8, 18, 36, 25].”
Instance Embedding Transfer to Unsupervised Video Object Segmentation (2018, methodology): “Classic work in video segmentation produced results using optical flow and shallow appearance models [20, 29, 22, 32, 11] while more recent methods typically pretrain the network Figure 1.”
Layered segmentation and optical flow estimation over time (2012, methodology): “For layer segmentation, we also compare our method to a state-of-the-art, hierarchical graph-based video segmentation algorithm [9], referred to”
Click Carving: Segmenting Objects in Video with Point Clicks (2016, methodology): “The problem has received substantial attention in recent years, with methods ranging from wholly unsupervised bottom-up approaches (Grundmann et al. 2010; Xu, Xiong, and Corso 2012), to propagation methods that exploit user input on the first frame (Ren and Malik 2007; Tsai, Flagg, and Rehg 2010; Fa”
FusionSeg: Learning to Combine Motion and Appearance for Fully Automatic Segmentation of Generic Objects in Videos (2017, methodology): “Our method also outperforms the semi-supervised HVS [1] method in 8 out of 14 cases.”
Track and Segment: An Iterative Unsupervised Approach for Video Object Proposals (2016, methodology): “The approach of [19] clusters pixels using bottom-up motion and appearance cues, and thus needs to generate many proposals to obtain good performance.”