Graph cut based image segmentation with connectivity priors

By Sara Alexandra Gomes Vicente, Vladimir Kolmogorov, Carsten Rother. 2008

454 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

Extreme Clicking for Efficient Object Annotation (2017, methodology): “To reduce annotation time many interactive segmentation techniques have been proposed, which require the user to input either a bounding box around the object [42, 51, 71], or scribbles [3, 16, 24, 26, 27, 42, 50, 64, 65, 72], or clicks [31, 69].”
Lifting Object Detection Datasets into 3D (2015, background): “This happens because the algorithm trusts equally all silhouettes.”
Image segmentation with a bounding box prior (2009, background): “Note, that the connectivity property is, in general, hard to incorporate into global optimization frameworks [15, 12].”
Generalized Connectivity Constraints for Spatio-temporal 3D Reconstruction (2014, methodology): “[26] use connectivity priors for a Markov random field segmentation.”
Automatic salient object segmentation based on context and shape prior (2011, methodology): “Our work differs from the previous works mentioned above mostly because we incorporated the generic knowledge of object into salient object segmentation.”
ImageNet Auto-Annotation with Segmentation Propagation (2014, background): “The pairwise potential is Analog to [52,10,70,59,12], this potential encourages smoothness by penalizing neighbouring pixels taking different labels.”
Y Qualification Main task Y Return N N Y User qualified ? N Feedback Pass ? Pass ? Annotator training Annotating images (2017, methodology): “To reduce annotation time many interactive segmentation techniques have been proposed, which require the user to input either a bounding box around the object [42, 51, 71], or scribbles [3, 16, 24, 26, 27, 42, 50, 64, 65, 72], or clicks [31, 69].”
Asymmetric Cuts: Joint Image Labeling and Partitioning (2014, background): “[33], but penalizes a weighted sum of cut edges which leads to the problem of shrinking bias [36], for which sophisticated countermeasures were developed.”
clDice - a Novel Topology-Preserving Loss Function for Tubular Structure Segmentation (2020, methodology): “For non-binary geometries, existing methods applied topology and connectivity constraints onto variational and Markov random field-based methods: tree shape priors for vessel segmentation [44], graph representation priors to natural images [2], higher-order cliques which connect superpixels [53] and”
Geodesic star convexity for interactive image segmentation (2010, background): “Notice how the starconvexity constraint helps to remove disconnected FG islands, and also to connect up FG islands into a single component. nectivity constraints for segmentation by Vicente et al. [26] and Nowozin et al. [15].”
Segmentation Propagation in ImageNet (2012, unknown): “Analog to [1,26,32,33,34], this potential encourages smoothness by penalizing neighboring pixels taking different labels.”
Reconstructing PASCAL VOC (2014, methodology): “We choose to formulate our reconstruction algorithm as a labelling problem, to motivate future extensions such as adding pairwise constraints between voxels or connectivity priors [26].”
MultiSeg: Semantically Meaningful, Scale-Diverse Segmentations From Minimal User Input (2019, background): “Various forms of priors have also been proposed to further improve the segmentation performance [10, 12, 35].”
Beat the MTurkers: Automatic Image Labeling from Weak 3D Supervision (2014, methodology): “[31] use DijkstraGC to impose object connectivity priors into segmentation.”
MILCut: A Sweeping Line Multiple Instance Learning Paradigm for Interactive Image Segmentation (2014, background): “While fully automated segmentation is arguably an intrinsically ambiguous problem [5] and manual segmentation is time-consuming to obtain, semi-automated (user-interactive) segmentation has demonstrated great practical importance and popularity [10, 32, 15, 9, 22, 29, 19, 8, 35, 27, 20].”