Laurent Jacques

Robust 1-Bit Compressive Sensing via Binary Stable Embeddings of Sparse Vectors

The compressive sensing (CS) framework aims to ease the burden on analog-to-digital converters (ADCs) by reducing the sampling rate required to acquire and stably recover sparse signals. Practical ADCs not only sample but also quantize each measurement to a finite number of bits; moreover, there is an inverse relationship between the achievable sampling rate and the bit depth. In this paper, we investigate an alternative CS approach that shifts the emphasis from the sampling rate to the number of bits per measurement. In particular, we explore the extreme case of 1-bit CS measurements, which capture just their sign. Our results come in two flavors. First, we consider ideal reconstruction from noiseless 1-bit measurements and provide a lower bound on the best achievable reconstruction error. We also demonstrate that i.i.d. random Gaussian matrices provide measurement mappings that, with overwhelming probability, achieve nearly optimal error decay. Next, we consider reconstruction robustness to measurement errors and noise and introduce the binary ε-stable embedding property, which characterizes the robustness of the measurement process to sign changes. We show that the same class of matrices that provide almost optimal noiseless performance also enable such a robust mapping. On the practical side, we introduce the binary iterative hard thresholding algorithm for signal reconstruction from 1-bit measurements that offers state-of-the-art performance.

Dequantizing Compressed Sensing: When Oversampling and Non-Gaussian Constraints Combine

In this paper, we study the problem of recovering sparse or compressible signals from uniformly quantized measurements. We present a new class of convex optimization programs, or decoders, coined Basis Pursuit DeQuantizer of moment p (BPDQp), that model the quantization distortion more faithfully than the commonly used Basis Pursuit DeNoise (BPDN) program. Our decoders proceed by minimizing the sparsity of the signal to be reconstructed subject to a data-fidelity constraint expressed in the ℓp-norm of the residual error for 2 ≤ p ≤ ∞. We show theoretically that, (i) the reconstruction error of these new decoders is bounded if the sensing matrix satisfies an extended Restricted Isometry Property involving the Iρ norm, and (ii), for Gaussian random matrices and uniformly quantized measurements, BPDQp performance exceeds that of BPDN by dividing the reconstruction error due to quantization by √(p + 1). This last effect happens with high probability when the number of measurements exceeds a value growing with p, i.e., in an oversampled situation compared to what is commonly required by BPDN = BPDQ2. To demonstrate the theoretical power of BPDQp, we report numerical simulations on signal and image reconstruction problems.

Compressed sensing imaging techniques for radio interferometry

Radio interferometry probes astrophysical signals through incomplete and noisy Fourier measurements. The theory of compressed sensing demonstrates that such measurements may actually suffice for accurate reconstruction of sparse or compressible signals. We propose new generic imaging techniques based on convex optimization for global minimization problems defined in this context. The versatility of the framework notably allows introduction of specific prior information on the signals, which offers the possibility of significant improvements of reconstruction relative to the standard local matching pursuit algorithm CLEAN used in radio astronomy. We illustrate the potential of the approach by studying reconstruction performances on simulations of two different kinds of signals observed with very generic interferometric configurations. The first kind is an intensity field of compact astrophysical objects. The second kind is the imprint of cosmic strings in the temperature field of the cosmic microwave background radiation, of particular interest for cosmology.

Wavelets on the sphere: implementation and approximations

We continue the analysis of the continuous wavelet transform on the 2-sphere, introduced in a previous paper. After a brief review of the transform, we define and discuss the notion of directional spherical wavelet, i.e., wavelets on the sphere that are sensitive to directions. Then we present a calculation method for data given on a regular spherical grid g. This technique, which uses the FFT, is based on the invariance of g under discrete rotations around the z axis preserving the phi sampling. Next, a numerical criterion is given for controlling the scale interval where the spherical wavelet transform makes sense, and examples are given, both academic and realistic. In a second part, we establish conditions under which the reconstruction formula holds in strong L-p sense, for 1 less than or equal to p < infinity. This opens the door to techniques for approximating functions on the sphere, by use of an approximate identity, obtained by a suitable dilation of the mother wavelet

Correspondence principle between spherical and Euclidean wavelets

Wavelets on the sphere are reintroduced and further developed independently of the original group theoretic formalism, in an equivalent, but more straightforward approach. These developments are motivated by the interest of the scale-space analysis of the cosmic microwave background (CMB) anisotropies on the sky. A new, self-consistent, and practical approach to the wavelet filtering on the sphere is developed. It is also established that the inverse stereographic projection of a wavelet on the plane (i.e. Euclidean wavelet) leads to a wavelet on the sphere (i.e. spherical wavelet). This new correspondence principle simplifies the construction of wavelets on the sphere and allows to transfer onto the sphere properties of wavelets on the plane. In that regard, we define and develop the notions of directionality and steerability of filters on the sphere. In the context of the CMB analysis, these notions are important tools for the identification of local directional features in the wavelet coefficients of the signal, and for their interpretation as possible signatures of non-gaussianity, statistical anisotropy, or foreground emission. But the generic results exposed may find numerous applications beyond cosmology and astrophysics. Comment: Version accepted in ApJ. Clarified discussion on the unicity of the wavelet formalism on the sphere and on the related correspondence principle. 14 pages, 8 figures, Revtex4 (emulateapj)

A panorama on multiscale geometric representations, intertwining spatial, directional and frequency selectivity

The richness of natural images makes the quest for optimal representations in image processing and computer vision challenging. The latter observation has not prevented the design of image representations, which trade off between efficiency and complexity, while achieving accurate rendering of smooth regions as well as reproducing faithful contours and textures. The most recent ones, proposed in the past decade, share a hybrid heritage highlighting the multiscale and oriented nature of edges and patterns in images. This paper presents a panorama of the aforementioned literature on decompositions in multiscale, multi-orientation bases or dictionaries. They typically exhibit redundancy to improve sparsity in the transformed domain and sometimes its invariance with respect to simple geometric deformations (translation, rotation). Oriented multiscale dictionaries extend traditional wavelet processing and may offer rotation invariance. Highly redundant dictionaries require specific algorithms to simplify the search for an efficient (sparse) representation. We also discuss the extension of multiscale geometric decompositions to non-Euclidean domains such as the sphere or arbitrary meshed surfaces. The etymology of panorama suggests an overview, based on a choice of partially overlapping “pictures”. We hope that this paper will contribute to the appreciation and apprehension of a stream of current research directions in image understanding.

CMOS compressed imaging by Random Convolution

We present a CMOS imager with built-in capability to perform Compressed Sensing coding by Random Convolution. It is achieved by a shift register set in a pseudo-random configuration. It acts as a convolutive filter on the imager focal plane, the current issued from each CMOS pixel undergoing a pseudo-random redirection controlled by each component of the filter sequence. A pseudo-random triggering of the ADC reading is finally applied to complete the acquisition model. The feasibility of the imager and its robustness under noise and non-linearities have been confirmed by computer simulations, as well as the reconstruction tools supporting the Compressed Sensing theory.

Diffeomorphic registration of images with variable contrast enhancement

Nonrigid image registration is widely used to estimate tissue deformations in highly deformable anatomies. Among the existing methods, nonparametric registration algorithms such as optical flow, or Demons, usually have the advantage of being fast and easy to use. Recently, a diffeomorphic version of the Demons algorithm was proposed. This provides the advantage of producing invertible displacement fields, which is a necessary condition for these to be physical. However, such methods are based on the matching of intensities and are not suitable for registering images with different contrast enhancement. In such cases, a registration method based on the local phase like the Morphons has to be used. In this paper, a diffeomorphic version of the Morphons registration method is proposed and compared to conventional Morphons, Demons, and diffeomorphic Demons. The method is validated in the context of radiotherapy for lung cancer patients on several 4D respiratory-correlated CT scans of the thorax with and without variable contrast enhancement.

Fast directional correlation on the sphere with steerable filters

A fast algorithm is developed for the directional correlation of scalar band-limited signals and band-limited steerable filters on the sphere. The asymptotic complexity associated to it through simple quadrature is of order O(L^5), where 2L stands for the square-root of the number of sampling points on the sphere, also setting a band limit L for the signals and filters considered. The filter steerability allows to compute the directional correlation uniquely in terms of direct and inverse scalar spherical harmonics transforms, which drive the overall asymptotic complexity. The separation of variables technique for the scalar spherical harmonics transform produces an O(L^3) algorithm independently of the pixelization. On equi-angular pixelizations, a sampling theorem introduced by Driscoll and Healy implies the exactness of the algorithm. The equi-angular and HEALPix implementations are compared in terms of memory requirements, computation times, and numerical stability. The computation times for the scalar transform, and hence for the directional correlation, of maps of several megapixels on the sphere (L~10^3) are reduced from years to tens of seconds in both implementations on a single standard computer. These generic results for the scale-space signal processing on the sphere are specifically developed in the perspective of the wavelet analysis of the cosmic microwave background (CMB) temperature (T) and polarization (E and B) maps of the WMAP and Planck experiments. As an illustration, we consider the computation of the wavelet coefficients of a simulated temperature map of several megapixels with the second Gaussian derivative wavelet. Comment: Version accepted in APJ. 14 pages, 2 figures, Revtex4 (emulateapj). Changes include (a) a presentation of the algorithm as directly built on blocks of standard spherical harmonics transforms, (b) a comparison between the HEALPix and equi-angular implementations

Sparsity Driven People Localization with a Heterogeneous Network of Cameras

This paper addresses the problem of localizing people in low and high density crowds with a network of heterogeneous cameras. The problem is recast as a linear inverse problem. It relies on deducing the discretized occupancy vector of people on the ground, from the noisy binary silhouettes observed as foreground pixels in each camera. This inverse problem is regularized by imposing a sparse occupancy vector, i.e., made of few non-zero elements, while a particular dictionary of silhouettes linearly maps these non-empty grid locations to the multiple silhouettes viewed by the cameras network. The proposed framework is (i) generic to any scene of people, i.e., people are located in low and high density crowds, (ii) scalable to any number of cameras and already working with a single camera, (iii) unconstrained by the scene surface to be monitored, and (iv) versatile with respect to the camera’s geometry, e.g., planar or omnidirectional.Qualitative and quantitative results are presented on the APIDIS and the PETS 2009 Benchmark datasets. The proposed algorithm successfully detects people occluding each other given severely degraded extracted features, while outperforming state-of-the-art people localization techniques.