Scattering spectra models for physics.

Physicists routinely need probabilistic models for a number of tasks such as parameter inference or the generation of new realizations of a field. Establishing such models for highly non-Gaussian fields is a challenge, especially when the number of samples is limited. In this paper, we introduce scattering spectra models for stationary fields and we show that they provide accurate and robust statistical descriptions of a wide range of fields encountered in physics. These models are based on covariances of scattering coefficients, i.e. wavelet decomposition of a field coupled with a pointwise modulus. After introducing useful dimension reductions taking advantage of the regularity of a field under rotation and scaling, we validate these models on various multiscale physical fields and demonstrate that they reproduce standard statistics, including spatial moments up to fourth order. The scattering spectra provide us with a low-dimensional structured representation that captures key properties encountered in a wide range of physical fields. These generic models can be used for data exploration, classification, parameter inference, symmetry detection, and component separation.

Solid harmonic wavelet scattering for predictions of molecule properties.

We present a machine learning algorithm for the prediction of molecule properties inspired by ideas from density functional theory (DFT). Using Gaussian-type orbital functions, we create surrogate electronic densities of the molecule from which we compute invariant "solid harmonic scattering coefficients" that account for different types of interactions at different scales. Multilinear regressions of various physical properties of molecules are computed from these invariant coefficients. Numerical experiments show that these regressions have near state-of-the-art performance, even with relatively few training examples. Predictions over small sets of scattering coefficients can reach a DFT precision while being interpretable.

Understanding deep convolutional networks.

Deep convolutional networks provide state-of-the-art classifications and regressions results over many high-dimensional problems. We review their architecture, which scatters data with a cascade of linear filter weights and nonlinearities. A mathematical framework is introduced to analyse their properties. Computations of invariants involve multiscale contractions with wavelets, the linearization of hierarchical symmetries and sparse separations. Applications are discussed.

Scattering transform for intrapartum fetal heart rate variability fractal analysis: a case-control study.

Intrapartum fetal heart rate monitoring, aiming at early acidosis detection, constitutes an important public health stake. Scattering transform is proposed here as a new tool to analyze intrapartum fetal heart rate (FHR) variability. It consists of a nonlinear extension of the underlying wavelet transform, that thus preserves its multiscale nature. Applied to an FHR signal database constructed in a French academic hospital, the scattering transform is shown to permit to efficiently measure scaling exponents characterizing the fractal properties of intrapartum FHR temporal dynamics, that relate not only to the sole covariance (correlation scaling exponent), but also to the full dependence structure of data (intermittency scaling exponent). Such exponents are found to satisfactorily discriminate temporal dynamics of healthy subjects (from that of nonhealthy ones) and to emphasize the role of the highest frequencies (around and above 1 Hz) in intrapartum FHR variability. This permits us to achieve satisfactory classification performance that improves on those obtained from the analysis of International Federation of Gynecology and Obstetrics (FIGO) criteria, notably by classifying as healthy a number of subjects that were incorrectly classified as nonhealthy by classical clinically used FIGO criteria. Combined to obstetrician annotations, these scaling exponents enable us to sketch a typology of these FIGO-false positive subjects. Also, they permit us to monitor the evolution along time of the intrapartum health status of the fetuses and to estimate an optimal detection time-frame.

Scattering transform for intrapartum fetal heart rate characterization and acidosis detection.

Early acidosis detection and asphyxia prediction in intrapartum fetal heart rate is of major concern. This contribution aims at assessing the potential of the Scattering Transform to characterize intrapartum fetal heart rate. Elaborating on discrete wavelet transform, the Scattering Transform performs a non linear and multiscale analysis, thus probing not only the covariance structure of data but also the full dependence structure. Applied to a real database constructed by a French public academic hospital, the Scattering Transform is shown to catch relevant features of intrapartum fetal heart rate time dynamics and to have a satisfactory ability to discriminate Normal subjects from Abnormal.

Invariant scattering convolution networks.

A wavelet scattering network computes a translation invariant image representation which is stable to deformations and preserves high-frequency information for classification. It cascades wavelet transform convolutions with nonlinear modulus and averaging operators. The first network layer outputs SIFT-type descriptors, whereas the next layers provide complementary invariant information that improves classification. The mathematical analysis of wavelet scattering networks explains important properties of deep convolution networks for classification. A scattering representation of stationary processes incorporates higher order moments and can thus discriminate textures having the same Fourier power spectrum. State-of-the-art classification results are obtained for handwritten digits and texture discrimination, with a Gaussian kernel SVM and a generative PCA classifier.

Solving inverse problems with piecewise linear estimators: from Gaussian mixture models to structured sparsity.

A general framework for solving image inverse problems with piecewise linear estimations is introduced in this paper. The approach is based on Gaussian mixture models, which are estimated via a maximum a posteriori expectation-maximization algorithm. A dual mathematical interpretation of the proposed framework with a structured sparse estimation is described, which shows that the resulting piecewise linear estimate stabilizes the estimation when compared with traditional sparse inverse problem techniques. We demonstrate that, in a number of image inverse problems, including interpolation, zooming, and deblurring of narrow kernels, the same simple and computationally efficient algorithm yields results in the same ballpark as that of the state of the art.

Super-resolution with sparse mixing estimators.

We introduce a class of inverse problem estimators computed by mixing adaptively a family of linear estimators corresponding to different priors. Sparse mixing weights are calculated over blocks of coefficients in a frame providing a sparse signal representation. They minimize an l1 norm taking into account the signal regularity in each block. Adaptive directional image interpolations are computed over a wavelet frame with an O(N logN) algorithm, providing state-of-the-art numerical results.

Sparse geometric image representations with bandelets.

This paper introduces a new class of bases, called bandelet bases, which decompose the image along multiscale vectors that are elongated in the direction of a geometric flow. This geometric flow indicates directions in which the image gray levels have regular variations. The image decomposition in a bandelet basis is implemented with a fast subband-filtering algorithm. Bandelet bases lead to optimal approximation rates for geometrically regular images. For image compression and noise removal applications, the geometric flow is optimized with fast algorithms so that the resulting bandelet basis produces minimum distortion. Comparisons are made with wavelet image compression and noise-removal algorithms.

Deconvolution by thresholding in mirror wavelet bases.

The deconvolution of signals is studied with thresholding estimators that decompose signals in an orthonormal basis and threshold the resulting coefficients. A general criterion is established to choose the orthonormal basis in order to minimize the estimation risk. Wavelet bases are highly sub-optimal to restore signals and images blurred by a low-pass filter whose transfer function vanishes at high frequencies. A new orthonormal basis called mirror wavelet basis is constructed to minimize the risk for such deconvolutions. An application to the restoration of satellite images is shown.