A Statistical Approach to Texture Classification from Single Images

By Manik Varma, Andrew Zisserman. 2004

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

Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations (2016, methodology): “In the first, we treat attributes independently from objects and train a classifier for each attribute, i.e. a classifier for red or a classifier for old, as in (Malisiewicz et al., 2008, Varma and Zisserman, 2005, Ferrari and Zisserman, 2007, Farhadi et al., 2009, Johnson et al., 2015).”
Framing Image Description as a Ranking Task: Data, Models and Evaluation Metrics (2013, methodology): “K T captures texture, represented by descriptors which capture edge information at different orientations centered on the pixel (Varma & Zisserman, 2005).”
Describing Textures in the Wild (2013, methodology): “The ﬁrst one is the use of the SIFT descriptors, originally developed for object matching [20], that are more distinctive that local descriptors popular in texture analysis such as ﬁlter banks [13, 19, 36], local intensity patterns [23], and patches [35].”
Material recognition in the wild with the Materials in Context Database (2014, background): “This led to research on the task of instance-level texture or material classification [15, 30], and an appreciation of the challenges of building features that are invariant to pose and illumination.”
Large scale visual recommendations from street fashion images (2014, methodology): “We performed extensive experiments with texture features such Gabor Wavelets [10], Textons [12], Dense-SIFT[2] and found them to be unstable for texture classiﬁ-cation in clothing.”
An empirical study of context in object detection (2009, methodology): “For class appearance, we compute Kmeans clustered L*a*b* color (K=128) and texton [40] (K=256) histograms, geometric context confidences [16] and the probability of background confidences (trained using [16] on LabelMe [33] examples), quantized to ten values.”
Fusing generic objectness and visual saliency for salient object detection (2011, methodology): “We exploit the fact that the detector is formed by a naı̈ve Bayes model where each cue is considered independently, and modify it by removing the saliency cue in all our experiments.”
Object categorization by learned universal visual dictionary (2005, methodology): “Specifically, we have borrowed the texton-based models developed in the context of texture recognition [9] [16] and extended by [1].”
A Statistical Approach to Material Classification Using Image Patch Exemplars (2009, methodology): “Later algorithms such as the BFH classiﬁer [9] and the VZ classiﬁer [52] have built on this paper and extended it to classify single images without compromising accuracy.”
Pairwise Rotation Invariant Co-Occurrence Local Binary Pattern (2014, background): “It plays a significant role in a wide range of applications that includes static and dynamic texture classification [1], [2], [3], [4], [5], object and scene recognition [6], [7], [8], [9], [10], face detection and recognition [11], image retrieval [12], [13], stereo correspondence, 3D…”
Selection of the Mars Science Laboratory Landing Site (2012, methodology): “A machine vision algorithm was used to segments images via pattern recognition into regions with similar textures (Ojala and Pietikainen 1999; Varma and Zisserman 2005; Cheng et al. 2003).”
Content-Based High-Resolution Remote Sensing Image Retrieval via Unsupervised Feature Learning and Collaborative Affinity Metric Fusion (2016, methodology): “LBP in [15] GLCM in [16] MR8 in [17] SIFT in [18] UCNN1 UCNN2 UCNN3 UCNN4”
Texture Feature Extraction Methods: A Survey (2019, background): “In 2005, Varma and Zisserman [237] proposed a statistical learning approach (VZ-MR8) : textures were modelled by the joint distribution of filter responses.”
Viewpoint Invariant Pedestrian Recognition with an Ensemble of Localized Features (2008, background): “Histograms have proven useful for an assortment of tasks including tracking [9], texture classification [10], and pedestrian detection [11].”
Performance evaluation of local colour invariants (2009, unknown): “Fur thermore , we will evaluate the hue - based SIFT descriptor of Ab del - Hakim and Farag [ 53 ] , termed hue - colour - SIFT , and the HSV - based SIFT descriptor of Bosch and Zisserman [ 54 ] , termed hsv - colour - SIFT .”
From BoW to CNN: Two Decades of Texture Representation for Texture Classification (2018, methodology): “Realizing the shortcomings of Leung and Malik’s method [114], Varma and Zisserman [236] attempted to improve the classiﬁcation of a single texture sample image under unknown imaging conditions, bypassing the registration step, instead learning 2D textons by aggregating ﬁlter responses over…”
Shape and Appearance Context Modeling (2007, background): “Bag-of-features approaches [20, 12, 14, 23, 26, 4], represent images, or portions of images, by a distribution/collection of, possibly densely computed, local descriptors, vector-quantized according to a predefined appearance dictionary (made of appearance labels).”
Discriminative Object Class Models of Appearance and Shape by Correlatons (2006, methodology): “Such methods model the appearance of each object class by learning the typical proportions of visual words - or textons (as originally introduced in the context of texture recognition [13, 19]).”
TextonBoost: Joint Appearance, Shape and Context Modeling for Multi-class Object Recognition and Segmentation (2006, methodology): “For this we utilize textons [17] which have been proven effective in categorizing materials [18] as well as generic object classes [4].”