FCDSN-DC: An Accurate and Lightweight Convolutional Neural Network for Stereo Estimation with Depth Completion

TL;DR

This paper introduces FCDSN-DC, a lightweight, fully-convolutional neural network that improves stereo estimation and depth completion accuracy while maintaining efficiency, suitable for diverse real-world scenes.

Contribution

The paper presents a novel, fully-convolutional deformable similarity network with depth completion that enhances feature extraction, similarity learning, and missing data filling in stereo depth estimation.

Findings

Achieves competitive results on Middlebury, KITTI, and ETH3D datasets.

Generalizes well across indoor and outdoor scenes without additional training.

Maintains high accuracy with a lightweight model suitable for real-world applications.

Abstract

We propose an accurate and lightweight convolutional neural network for stereo estimation with depth completion. We name this method fully-convolutional deformable similarity network with depth completion (FCDSN-DC). This method extends FC-DCNN by improving the feature extractor, adding a network structure for training highly accurate similarity functions and a network structure for filling inconsistent disparity estimates. The whole method consists of three parts. The first part consists of fully-convolutional densely connected layers that computes expressive features of rectified image pairs. The second part of our network learns highly accurate similarity functions between this learned features. It consists of densely-connected convolution layers with a deformable convolution block at the end to further improve the accuracy of the results. After this step an initial disparity map is created and the left-right consistency check is performed in order to remove inconsistent points. The last part of the network then uses this input together with the corresponding left RGB image in order to train a network that fills in the missing measurements. Consistent depth estimations are gathered around invalid points and are parsed together with the RGB points into a shallow CNN network structure in order to recover the missing values. We evaluate our method on challenging real world indoor and outdoor scenes, in particular Middlebury, KITTI and ETH3D were it produces competitive results. We furthermore show that this method generalizes well and is well suited for many applications without the need of further training. The code of our full framework is available at: https://github.com/thedodo/FCDSN-DC