Contour and Texture Analysis for Image Segmentation

By Jitendra Malik, Serge J. Belongie, Thomas K. Leung, Jianbo Shi. 2001

1,370 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

Fast Edge Detection Using Structured Forests (2014, neutral): “segmentation [1], [33], and active contours [26].”
On the Binding Problem in Artificial Neural Networks (2020, background): “As we have argued, a hardwired similarity function (e.g. as in Shi and Malik, 2000; Malik et al., 2001) has little chance at facilitating the required flexibility, although different initial seedings of the clustering may still account for multiple different groupings (i.e. multistability).”
DeepEdge: A multi-scale bifurcated deep network for top-down contour detection (2014, methodology): “Due to locality of our method, it may be beneficial to apply spectral methods [21, 30] or conditional random fields [27] on top of our method to further improve its performance.”
Crisp Boundary Detection Using Pointwise Mutual Information (2014, background): “[1] embeds the local Pb measure into a spectral clustering framework [21,22].”
Image Segmentation with Ratio Cut (2003, methodology): “This lack of bias is shared by some other approaches to graphbased image segmentation, for example, the method of Jermyn and Ishikawa [4], [5] when the intensity gradient is consistently oriented inward or outward.”
Spectral grouping using the Nystrom method (2004, methodology): “In this case, each pixel in the image is associated with the nearest element in a small alphabet of prototypical linear filter responses using vector-quantitization (see Malik et al. [18]).”
Coupled information-theoretic encoding for face photo-sketch recognition (2011, methodology): “Vector quantization was widely used to create discrete image representations, such as textons [20] and visual words [24], for object recognition and face recognition.”
Segmentation of Multivariate Mixed Data via Lossy Data Coding and Compression (2007, background): “More visually appealing segmentations might be obtained with more sophisticated features (e.g., filterbanks [ 32 ], [33]).”
Pulling Things out of Perspective (2014, methodology): “Following the methodology of the multi-feature [13] extension of TextonBoost [24], the dense features, namely texton [18], SIFT [17], local quantized ternary patters [12] and self-similarity features [21], are extracted for each pixel in each image in a pyramid α ∗ I .”
Contour Detection and Hierarchical Image Segmentation (2011, methodology): “Spectral graph theory [48] and, in particular, the Normalized Cuts criterion [45], [46] provides a way of integrating global image information into the grouping 900 IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL.”
Shape-Texture Debiased Neural Network Training (2020, background): “, textons (Malik et al., 2001)), for object recognition.”
Semantic3D.net: A new Large-scale Point Cloud Classification Benchmark (2017, methodology): “ for a variety of tasks (e.g., (Montoya et al., 2014, Ladicky´ et al., 2014)) and was available in its original implementation. The method works as follows: four different types of features – texton (Malik et al., 2001), SIFT (Lowe, 2004), local quantized ternary patters (Hussain and Triggs, 2012) a”
From contours to regions: An empirical evaluation (2009, background): “A broad family of approaches to segmentation involve integrating features such as brightness, color, or texture over local image patches and then clustering those features based on, e.g., fitting mixture models [5] [47], mode-finding [10], or graph partitioning [42] [24] [44] [15].”
Semantic texton forests for image categorization and segmentation (2008, background): “Textons [11, 17, 31] and visual words [28] have proven powerful discrete image representations for categorization and segmentation [4, 27, 34, 37].”
Multiclass spectral clustering (2003, methodology): “Images are first convolved with oriented filter pairs to extract the magnitude of edge responses OE [6].”