Abstract:Scene flow estimation is the task to predict the point-wise or pixel-wise 3-D displacement vector between two consecutive frames of point clouds or images, which has important application in fields such as service robots and autonomous driving. Although many previous works have explored greatly on scene flow estimation based on point clouds, there are two problems that have not been noticed or well solved before: 1) points of adjacent frames in repetitive patterns tend to be wrongly associated due to similar spatial structure in their neighborhoods and 2) scene flow between adjacent frames of point clouds with long-distance movement tends to be inaccurately estimated. To solve the first problem, a novel context-aware set convolution layer is proposed in this article to exploit contextual structure information of Euclidean space and learn soft aggregation weights for local point features. This design is inspired by human perception of contextual structure information during scene understanding with repetitive patterns. For the second problem, an explicit residual flow learning structure is proposed in the residual flow refinement layer to cope with long-distance movement. The experiments and ablation study on FlyingThings3D and Karlsruhe Institute of Technology and Toyota Technological Institute (KITTI) scene flow datasets demonstrate the effectiveness of each proposed component. The qualitative results show that the problems of ambiguous interframe association and long-distance movement estimation are well handled. Quantitative results on both FlyingThings3D and KITTI scene flow datasets show that the proposed method achieves state-of-the-art performance, surpassing all previous works to the best of our knowledge by at least 26.22% on FlyingThings3D dataset and 25.42% on KITTI scene flow dataset for 3-D endpoint error (EPE3D) metric. The source codes are available at: https://github.com/IRMVLab/CARFlow.

Active Scene Flow Estimation for Autonomous Driving via Real-Time Scene Prediction and Optimal Decision

Unsupervised Learning of Scene Flow Estimation Fusing with Local Rigidity.

Learning Graph-Based Representations for Scene Flow Estimation

SeFlow: A Self-Supervised Scene Flow Method in Autonomous Driving

Residual 3D Scene Flow Learning with Context-Aware Feature Extraction

Self-Supervised Multi-Frame Neural Scene Flow

Moving Object Detection Based on 3D Scene Flow for Autonomous Vehicles

Scalable Scene Flow from Point Clouds in the Real World

Re-Evaluating LiDAR Scene Flow for Autonomous Driving

Let Occ Flow: Self-Supervised 3D Occupancy Flow Prediction

What Matters for 3D Scene Flow Network.

Self-Supervised Scene Flow Estimation with 4-D Automotive Radar

Motion Estimation Using Region-Level Segmentation and Extended Kalman Filter for Autonomous Driving

SFGAN: Unsupervised Generative Adversarial Learning of 3D Scene Flow from the 3D Scene Self

Unsupervised Learning of 3D Scene Flow with 3D Odometry Assistance

Interactive Prediction and Decision-Making for Autonomous Vehicles: Online Active Learning with Traffic Entropy Minimization

3DSFLabelling: Boosting 3D Scene Flow Estimation by Pseudo Auto-labelling

AdaOcc: Adaptive Forward View Transformation and Flow Modeling for 3D Occupancy and Flow Prediction

ICP-Flow: LiDAR Scene Flow Estimation with ICP

Dynamic obstacle detection method based on U–V disparity and residual optical flow for autonomous driving

3-D Scene Flow Estimation on Pseudo-LiDAR: Bridging the Gap on Estimating Point Motion