Temp-Frustum Net: 3D Object Detection with Temporal Fusion

TL;DR

Temp-Frustum Net introduces a temporal fusion module that leverages information from previous frames to improve 3D object detection robustness against noise, occlusion, and sparsity in autonomous driving scenarios.

Contribution

It proposes a novel Temporal Fusion Module (TFM) integrated with a frustum network to enhance 3D detection by utilizing temporal information, outperforming frame-by-frame methods.

Findings

Achieves ~6% improvement on Car detection

Achieves ~4% improvement on Pedestrian detection

Achieves ~6% improvement on Cyclist detection

Abstract

3D object detection is a core component of automated driving systems. State-of-the-art methods fuse RGB imagery and LiDAR point cloud data frame-by-frame for 3D bounding box regression. However, frame-by-frame 3D object detection suffers from noise, field-of-view obstruction, and sparsity. We propose a novel Temporal Fusion Module (TFM) to use information from previous time-steps to mitigate these problems. First, a state-of-the-art frustum network extracts point cloud features from raw RGB and LiDAR point cloud data frame-by-frame. Then, our TFM module fuses these features with a recurrent neural network. As a result, 3D object detection becomes robust against single frame failures and transient occlusions. Experiments on the KITTI object tracking dataset show the efficiency of the proposed TFM, where we obtain ~6%, ~4%, and ~6% improvements on Car, Pedestrian, and Cyclist classes, respectively, compared to frame-by-frame baselines. Furthermore, ablation studies reinforce that the subject of improvement is temporal fusion and show the effects of different placements of TFM in the object detection pipeline. Our code is open-source and available at https://github.com/emecercelik/Temp-Frustum-Net.git.