Rigid-Motion Scattering for Texture Classification

By Laurent Sifre, Stéphane Mallat. 2014

422 citations recorded by the source corpus.

A rigid-motion scattering computes adaptive invariants along translations and rotations, with a deep convolutional network. Convolutions are calculated on the rigid-motion group, with wavelets defined on the translation and rotation variables. It preserves joint rotation and translation information, while providing global invariants at any desired scale. Texture classification is studied, through the characterization of stationary processes from a single realization. State-of-the-art results are obtained on multiple texture data bases, with important rotation and scaling variabilities.

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

Aggregated Residual Transformations for Deep Neural Networks (2016, background): “Channel-wise convolutions are part of the separable convolutions in [35].”
On Buggy Resizing Libraries and Surprising Subtleties in FID Calculation (2021, neutral): “For linear filters, this corresponds to a “depthwise convolution”, using deep learning parlance [31,68].”
General E(2)-Equivariant Steerable CNNs (2019, methodology): “As a result, we are able to implement a wide range of equivariant models, covering regular GCNNs [6, 7, 8, 9, 10, 11], classical Steerable CNNs [1], Harmonic Networks [12], gated Harmonic Networks [2], Vector Field Networks [13], Scattering Transforms [14, 15, 16, 17, 18] and entirely new architectu”
Quartznet: Deep Automatic Speech Recognition with 1D Time-Channel Separable Convolutions (2019, methodology): “Our approach is directly related to MobileNets [7, 8] and Xception [9], which uses depthwise separable convolutions [10, 11].”
In Defense of Pre-Trained ImageNet Architectures for Real-Time Semantic Segmentation of Road-Driving Images (2019, background): “Depthwise separable convolutions [33, 36] decrease computational complexity by splitting a regular convolution in two.”
A Mathematical Theory of Deep Convolutional Neural Networks for Feature Extraction (2015, background): “…step [51], and then feeds the result into a trainable classiﬁer such as, e.g., a SVM. State-of-the-art results for scattering networks were reported for various classiﬁcation tasks such as handwritten digit recognition [23], texture discrimination [23], [24], and musical genre classiﬁcation [25].”
A Survey of Convolutional Neural Networks: Analysis, Applications, and Prospects (2020, background): “Moreover, there exist a variety of awesome convolutions, such as Separable convolutions [24], [25], [26], [27], [28], group convolutions [11], [29], [30], [31] and multi-dimensional convolutions, which are discussed in Section 3 and Section 5.”
Twins: Revisiting the Design of Spatial Attention in Vision Transformers (2021, background): “As an extreme case, depthwise convolutions [12, 36] use the number of groups that is”
Interleaved Group Convolutions for Deep Neural Networks (2017, methodology): “In the channel extent, group convolutions [42, 40] and channel-wise convolutions or separable filters [28, 4, 14], have been studied.”
On Aliased Resizing and Surprising Subtleties in GAN Evaluation (2022, methodology): “For linear filters, this corresponds to a “depth-wise convolution”, using deep learning parlance [27, 62].”
Machine Learning at the Network Edge: A Survey (2019, methodology): “Depthwise separable convolutions were introduced by Sifre and Mallat [96].”
Review of Image Classification Algorithms Based on Convolutional Neural Networks (2021, methodology): “Secondly, for the depthwise convolution [127] often cannot fully utilize modern accelerators.”