3D scattering transforms for disease classification in neuroimaging

Abstract

Classifying neurodegenerative brain diseases in MRI aims at correctly assigning discrete labels to MRI scans. Such labels usually refer to a diagnostic decision a learner infers based on what it has learned from a training sample of MRI scans. Classification from MRI voxels separately typically does not provide independent evidence towards or against a class; the information relevant for classification is only present in the form of complicated multivariate patterns (or “features”). Deep learning solves this problem by learning a sequence of non-linear transformations that result in feature representations that are better suited to classification. Such learned features have been shown to drastically outperform hand-engineered features in computer vision and audio analysis domains. However, applying the deep learning approach to the task of MRI classification is extremely challenging, because it requires a very large amount of data which is currently not available. We propose to instead use a three dimensional scattering transform, which resembles a deep convolutional neural network but has no learnable parameters. Furthermore, the scattering transform linearizes diffeomorphisms (due to e.g. residual anatomical variability in MRI scans), making the different disease states more easily separable using a linear classifier. In experiments on brain morphometry in Alzheimer's disease, and on white matter microstructural damage in HIV, scattering representations are shown to be highly effective for the task of disease classification. For instance, in semi-supervised learning of progressive versus stable MCI, we reach an accuracy of 82.7%. We also present a visualization method to highlight areas that provide evidence for or against a certain class, both on an individual and group level.

Wavelet scattering networks in deep learning for discovering protein markers in a cohort of Swedish rectal cancer patients

“…Among various convolutional network types, the wavelet‐scattering transform 31 , 32 , 33 has appeared as a topic of growing interest within the signal processing and machine learning communities. This approach continues finding applications in diverse fields, including but not limited to neural disease classification, 34 authentication of artworks, 35 predictive indoor fingerprinting‐based localization, 36 ECG beat classification, 37 classification of alcoholic EEG signals, 38 and magnetohydrodynamic simulations for pattern analysis. 39 …”

“…30 Among various convolutional network types, the wavelet-scattering transform [31][32][33] has appeared as a topic of growing interest within the signal processing and machine learning communities. This approach continues finding applications in diverse fields, including but not limited to neural disease classification, 34 authentication of artworks, 35 predictive indoor fingerprinting-based localization, 36 ECG beat classification, 37 classification of alcoholic EEG signals, 38 and magnetohydrodynamic simulations for pattern analysis. 39 Based on the successful applications of convolutional methods in CNNs and wavelet scattering, this paper attempts, for the first time, to combine these two types of convolution-based operations for extracting strongly differentiable features of protein-expressed IHC images.…”

Wavelet Scattering Networks of Deep-Learning Features for Robust Discovery of Protein Markers in Rectal Cancer

“…As another type of convolutional networks, the wavelet scattering transform [31][32][33] is gaining increasing attention from the community of signal processing and machine learning with its applications for pattern analysis, such as neural disease classification 34 , authentication of art works 35 , predicting indoor fingerprinting-based localization 36 , classification of ECG beats 37 , classification of alcoholic EEG signals 38 , and classification of magnetohydrodynamic simulations 39 .…”

Kymatio: Scattering Transforms in Python

“…It is defined as a convolutional network whose filters are fixed to be wavelet and lowpass averaging filters coupled with modulus nonlinearities. It has many favorable theoretical properties (Mallat, 2012;Bruna et al, 2015;Waldspurger, 2017) and enjoys considerable success as a powerful tool in modern signal processing (Adel et al, 2017;Bruna and Mallat, 2013;Andén and Mallat, 2014;Chudáček et al, 2014;Sifre and Mallat, 2013;Eickenberg et al, 2017). It is also effective in combination with modern representation learning approaches (Oyallon et al, 2018;Sainath et al, 2014;Zeghidour et al, 2016).…”