Deep Convolutional Compressed Sensing for LiDAR Depth Completion

TL;DR

This paper introduces a deep convolutional compressed sensing approach using ADNNs for LiDAR depth completion, achieving superior results with a lightweight model.

Contribution

It presents a novel deep recurrent auto-encoder architecture that efficiently extracts multi-level sparse codes for dense depth map estimation from sparse LiDAR data.

Findings

Outperforms previous methods with only 1800 parameters

Uses a two-layer network to achieve state-of-the-art results

Demonstrates efficiency and effectiveness of compressed sensing in depth completion

Abstract

In this paper we consider the problem of estimating a dense depth map from a set of sparse LiDAR points. We use techniques from compressed sensing and the recently developed Alternating Direction Neural Networks (ADNNs) to create a deep recurrent auto-encoder for this task. Our architecture internally performs an algorithm for extracting multi-level convolutional sparse codes from the input which are then used to make a prediction. Our results demonstrate that with only two layers and 1800 parameters we are able to out perform all previously published results, including deep networks with orders of magnitude more parameters.