adribesson/icassp2018-pulse-streams

Pulse-stream models in time-of-flight imaging

Pulse-stream models in time-of-flight imaging

Adrien Besson, Dimitris Perdios, Marcel Arditi, Yves Wiaux and Jean-Philippe Thiran. submitted to IEEE ICASSP 2018

Introduction

This repository contains all the code to reproduce the results of the paper: Pulse-stream models in time-of-flight imaging.

Abstract:

This paper considers the problem of reconstructing raw signals from random projections in the context of time-of-flight imaging with an array of sensors. It presents a new signal model, coined as multichannel pulse-stream model, which exploits pulse-stream models and accounts for additional structure induced by inter-sensor dependencies. We propose a sampling theorem and a reconstruction algorithm, based on l1-minimization, for signals belonging to such a model. We show the benefits of the proposed approach by means of numerical simulations and on a real non-destructive-evaluation application.

Citation (BibTex):

If you are using this code, please cite the following paper: A. G. J. Besson, D. Perdios, Y. Wiaux and J.-P. Thiran. PULSE-STREAM MODELS IN TIME-OF-FLIGHT IMAGING submitted to 2018 IEEE International Conference on Acoustics, Speech and Signal Processing.

System Requirements:

This software has been tested on Linux 64-bit system (Ubuntu 16-04, Mint distribution) and on Windows 10 system.

Prerequisites

1. MATLAB (Tested on R2017a)

Running the code:

1. Download and unzip ICASSP2018-pulse-streams-master.zip
2. Open MATLAB and navigate in the folder ICASSP2018-pulse-streams-master
3. If you want to reproduce the Figures of the paper, run the script 'Script_reproduce_figure_ICASSP.m'
4. If you want to regenerate:
   • The results of the noiseless experiments (used to generate Figure 2), run the script 'Script_reproduce_noiseless_experiment.m'
   • The results of the noisy experiments (used to generate Figure 3), run the script 'Script_reproduce_noisy_experiment.m'
   • The results of the noisy experiment (used to generate Figure 4), run the script 'Script_reproduce_invivo_experiment.m'
5. If you are interested in testing the least-square solution of the problem, explained in Section 2.4.1 of the paper, for the noiseless experiment, you can run the script 'Script_least_squares_noiseless_experiment.m'

Contact:

Adrien Besson, ([email protected])

icassp2018-pulse-streams's Issues

Use of a Gaussian pulse for the simulation

Gaussian pulse for the impulse response rather than a sine wave

Change the matrix Psi with a 1D conv

Use a 1D conv with the pulse instead of a circulant matrix

Scripts for carotid reconstruction need to be cleaned

The script used to generate the results for the carotid experiment need to be cleaned

Allow to display the results when regenerating the experiments

Right now, when the experiments are regenerated, the only option is to store the results

Data upload

Data have to be uploaded -> which format?

Model-based iterative hard threshlding algorithm

Model-based IHT rather than ADMM should be implemented to fit model-based framework