Representing and Recognizing the Visual Appearance of Materials using Three-dimensional Textons

By Thomas K. Leung, Jitendra Malik. 2001

1,777 citations recorded by the source corpus.

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

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

Going deeper with convolutions (2014, methodology): “After estimating blur kernels, we apply the non-blind deconvolution method of Zoran and Weiss [6] to recover the whole latent images.”
Making Convolutional Networks Shift-Invariant Again (2019, background): “Shift-equivariance can be recovered if features are extracted densely, for example tex-tons (Leung & Malik, 2001), the Stationary Wavelet Trans-form (Fowler, 2005), and DenseSIFT (Vedaldi & Fulkerson, 2008).”
Describing Textures in the Wild (2013, methodology): “The most common coding method is to vector quantize the descriptors using an algorithm such as K -means [19], resulting in the so-called bag-of-visual-words (BoVW) representation [8].”
Understanding deep image representations by inverting them (2014, background): “Most image understanding and computer vision methods build on image representations such as textons [17], histogram of oriented gradients (SIFT [20] and HOG [4]), bag of visual words [3][27], sparse [37] and local coding [34], super vector coding [40], VLAD [10], Fisher Vectors [23], and, lately, de”
Learning Sight from Sound: Ambient Sound Provides Supervision for Visual Learning (2017, positive): “To study this, we replaced our sound texture features with (512-dimensional) visual texton histograms (Leung and Malik 2001), using the parameters from Xiao et al.”
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.”
Understanding Image Representations by Measuring Their Equivariance and Equivalence (2014, background): “Notable examples include textons (Leung and Malik 2001), histogram of oriented gradients (SIFT Lowe 2004) and HOG Dalal and Triggs 2005), bag of visual words (Csurka et al.”
Bilinear CNN Models for Fine-Grained Visual Recognition (2015, neutral): “showed boosting B-CNNs offers consistent improvements on ﬁne-grained tasks. Texture representations and second-order features. Texture representations have been widely studied for decades. Early work [35] represents the texture by computing the statistics of linear ﬁlter-bank responses (e.g., wavele”
Visualizing Deep Convolutional Neural Networks Using Natural Pre-images (2015, background): “Notable examples of representations include textons [22], histogram of oriented gradients (SIFT [25] and HOG [5]), bag of visual words [4][36], sparse [46] and local coding [43], super vector coding [51], VLAD [15], Fisher Vectors [31], and, lately,”
Unsupervised Discovery of Mid-Level Discriminative Patches (2012, background): “At a slightly higher level, combining such patches together, typically by clustering and histogramming, allows one to represent texture information (e.g., textons [9], dense bag-of-words [2], etc).”
Compact Bilinear Pooling (2015, methodology): “Leung and Malik used vector quantization in the Bag of Visual Words (BoVW) framework [20] initially used for texture classification, but later adopted for other visual tasks.”
Context Encoding for Semantic Segmentation (2018, methodology): “For a given input image, hand-engineered features are densely extracted using SIFT [37] or filter bank responses [30,46].”
Ambient Sound Provides Supervision for Visual Learning (2016, methodology): “To study this, we replaced our sound texture features with (512-dimensional) visual texton histograms [24], using the parameters from [37], and we used them to train a variation of our Clustering model.”
YOLACT: Real-Time Instance Segmentation (2019, methodology): “sets (13.5 fps on 5502 px images; see Table2c). Prototypes Learning prototypes (aka vocabulary or codebook) has been extensively explored in computer vision. Classical representations include textons [21] and visual words [37], with advances made via sparsity and locality priors [42,41,44]. Others h”
Oriented edge forests for boundary detection (2014, background): “These statistics can be captured in a variety of ways, e.g. via textons [15], sparse coding [22], or measures of self-similarity [14].”
The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS) (2015, methodology): “To characterize the tumor surface variation, we employ our novel texture features such as fractal PTPSA [107], and mBm [108] as well as classical textons [109] as the second set of features.”
Affordance Prediction via Learned Object Attributes (2011, methodology): “For texture features we extract a set of ﬁlter responses to create textons as described in [27].”
Domain-Specific Image Captioning (2014, methodology): “T EXTURE : Textons (Leung and Malik, 2001) are computed by convolving images with Gabor ﬁlters at multiple orientations and scales, then sampling the outputs at random locations.”