A Statistical Approach to Material Classification Using Image Patch Exemplars

By Manik Varma, Andrew Zisserman. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009

668 citations recorded by the source corpus.

Subfields: Atomic and Molecular Physics, and Optics, Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/2a65066f13a62d6f7e551a899d7f8a2fed063765

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Atomic and Molecular Physics, and Optics

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Texture Feature Extraction Methods: A Survey (2019, background): “Later, the same authors proposed another texton dictionarybased algorithm (VZ-Joint): instead of applying filter banks, Varma and Zisserman [238] proposed an alternative image patch representation based on the joint distribution of pixel intensities in a neighborhood.”
Rotation invariant co-occurrence features based on digital circles and discrete Fourier transform (2014, methodology): “The absence of tuning parameters, as well as the ease of implementation and other desirable asymptotic properties, support and explain the wide use of this method for comparative purposes (see for example Varma and Zisserman 2009; Guo et al. 2010; Crosier and Griffin 2010; Kandaswamy et al. 2011).”
Median Robust Extended Local Binary Pattern for Texture Classification (2016, methodology): “We demonstrate the performance of our approach with three different problems of robust texture classiﬁcation by conducting extensive experiments on a number of publicly available datasets, summarized in Table I, derived from the four most commonly used texture sources: Outex [48], CUReT [41], UMD [4”
Pairwise Rotation Invariant Co-Occurrence Local Binary Pattern (2014, methodology): “We use the same subset of images as [43], [44], which contains 61 texture classes with 92 images for each class.”
Physics 101: Learning Physical Object Properties from Unlabeled Videos (2016, methodology): “Our incorporation of physical laws into our model distinguish our framework from previous methods introduced in the computer vision and robotics community for predicting physical interactions or properties of objects for various purposes [4, 5, 9, 12, 16, 21, 22, 26].”
A Completed Modeling of Local Binary Pattern Operator for Texture Classification (2010, positive): “Later, Varma and Zisserman [7] [8] proposed another texton based algorithm by using the image local patch to represent features directly.”
An Analysis of Single-Layer Networks in Unsupervised Feature Learning (2011, methodology): “In particular, while clustering raw data leads to centroids consistent with those in [6] and [29], we see that clustering whitened data yields sharply localized filters that are very similar to those learned by the other algorithms .”
Scattering Representations for Recognition (2013, methodology): “The lowest published classification errors obtained on this dataset are 2% for Markov Random Fields [VZ09], 1.”
An Efficient Algorithm for Fractal Analysis of Textures (2012, methodology): “The two remaining datasets, KTH-TIPS [6] and Textured Surfaces [7] are publicly available and widely employed to evaluate texture analysis methods [8], [9], [10].”
Exploring features in a Bayesian framework for material recognition (2010, methodology): “[19]), it is significantly higher than the state of the art (23.8%, [ 30 ]).”
I Have Seen Enough: Transferring Parts Across Categories (2016, background): “This can be done by modeling semantic relatedness [38, 39], or by transferring other knowledge such as materials [5, 54], segments [45], parts [10, 49] or attributes [11, 29].”
The Synthesizability of Texture Examples (2014, background): “Features for material classification include statistics of filter responses [20, 28, 34], joint intensity distributions within a compact neighborhood [36, 22], geometric features over topographic maps [38], and high-level semantic attributes [29].”