From contours to regions: An empirical evaluation

By Pablo Arbeláez, Michael Maire, Charless C. Fowlkes, Jitendra Malik. 2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009

522 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/779dea9b105d8f48c9bc86f479b91f9c1d3c5963

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

Constrained parametric min-cuts for automatic object segmentation (2010, methodology): “[2], which is arguably the state of the art method for low-level segmentation.”
Object proposal by multi-branch hierarchical segmentation (2015, methodology): “The method in [4, 3, 5] gives merge scores using global contour detection result; the methods in [12, 26, 23] heuristically define merging strategies based on appearance similarity and segment’s size; the methods in [24, 27] trains a cascade of linear classifiers and uses classifiers’ decision value”
Layered object detection for multi-class segmentation (2010, methodology): “We show an example superpixel grouping from [1] tuned to return roughly 200 superpixels.”
Contour Detection and Hierarchical Image Segmentation (2011, methodology): “Paired with our gPb contour detector as input, our hierarchical segmentation algorithm gPb-owt-ucm [4] produces regions whose boundaries match ground truth better than those produced by other methods [7], [29], [30], [31], [32], [33], [34], [35].”
Segmentation of Natural Images by Texture and Boundary Compression (2010, unknown): “…whether these problems are unique to our method or are more universal, we have investigated similar problematic cases with the other methods reported here (Arbelaez et al. (2009); Comanicu & Meer (2002); Cour et al. (2005); Felzenszwalb & Huttenlocher (2004); Tu & Zhu (2002); Yang et al. (2008)).”
Layered Object Models for Image Segmentation (2012, methodology): “An example superpixel grouping from [33] tuned to return roughly 200 superpixels.”
Incorporating visual adjectives for image classification (2016, methodology): “To capture shape features, we compute boundary responses with Ultrametric Contour Map (UCM) [43], which is a closed contour extraction method based on a preliminary edge operator [44].”
Image Segmentation with Cascaded Hierarchical Models and Logistic Disjunctive Normal Networks (2013, methodology): “Since the task is detecting the boundary of cells, we compared our method with two general boundary detection methods, gPb-OWT-UCM (global probability of boundary followed by the oriented watershed transform and ultrametric contour maps) [1] and boosted edge learning (BEL) [9].”
Hierarchical Part Matching for Fine-Grained Visual Categorization (2013, methodology): “For this, we first calculate the Ultrametric Contour Map (UCM) [1], which generates closed contours with hierarchically decreasing boundary intensities, and cuts the image into smaller and smaller regions.”
Large displacement optical flow (2009, unknown): “We also integrate the boundary map g(x) from [1] (see Fig.”
Detecting People Using Mutually Consistent Poselet Activations (2010, methodology): “We use the state-of-the-art boundary predictor from [18] to obtain an edge map f : R(2) → [0, 1] of the image.”
Multi-task low-rank affinity pursuit for image segmentation (2011, methodology): “To quantitatively evaluate the performance of various methods, three metrics for comparing pairs of image segmentations are used: the variation of information (VOI) [23], the probabilistic rand index (PRI) [26] and the segmentation covering rate (CR) [2].”
Learning full pairwise affinities for spectral segmentation (2010, methodology): “The main research directions for this include mode-seeking [1], [2], deterministic annealing [3], stochastic clustering [4], [5], mixture model [6] [7], rate distortion [8], graph-based model [9], [10], [11], contourbased model [12], [13], and other variational methods [14], [15].”
Efficient, high-quality image contour detection (2009, methodology): “Our contours could be integrated into a method which produces image segments, such as [1], which can be more natural in some applications, such as object recognition [8].”
Recognition using regions (2009, methodology): “Regions are collected from all nodes of a region tree generated by [2].”
Object Recognition by Sequential Figure-Ground Ranking (2011, methodology): “We also both trained and tested with segments from Arbelaez et al. (2009) (OWT-UCM Masks)”
Coarse-to-Fine Description for Fine-Grained Visual Categorization (2016, methodology): “Similar to [10], E is constructed based on the Ultrametric Contour Map (UCM) [41].”
Hand detection using multiple proposals (2011, unknown): “Thus, the image is first split into super-pixels [3], then for a bounding box, the skin fraction (α3) is computed for the biggest super-pixel within it.”