SceneEDNet: A Deep Learning Approach for Scene Flow Estimation

TL;DR

This paper introduces SceneEDNet, a fully convolutional neural network designed for direct scene flow estimation from RGB-D videos, leveraging deep learning to improve accuracy and efficiency in robotics applications.

Contribution

It presents the first deep learning approach with an encoder-decoder architecture for direct scene flow estimation from stereo image sequences.

Findings

Effective estimation of 3D motion vectors from stereo images.

Application on large datasets yields meaningful scene flow results.

Advances in deep learning improve scene flow estimation accuracy.

Abstract

Estimating scene flow in RGB-D videos is attracting much interest of the computer vision researchers, due to its potential applications in robotics. The state-of-the-art techniques for scene flow estimation, typically rely on the knowledge of scene structure of the frame and the correspondence between frames. However, with the increasing amount of RGB-D data captured from sophisticated sensors like Microsoft Kinect, and the recent advances in the area of sophisticated deep learning techniques, introduction of an efficient deep learning technique for scene flow estimation, is becoming important. This paper introduces a first effort to apply a deep learning method for direct estimation of scene flow by presenting a fully convolutional neural network with an encoder-decoder (ED) architecture. The proposed network SceneEDNet involves estimation of three dimensional motion vectors of all the scene points from sequence of stereo images. The training for direct estimation of scene flow is done using consecutive pairs of stereo images and corresponding scene flow ground truth. The proposed architecture is applied on a huge dataset and provides meaningful results.