Frustum ConvNet: Sliding Frustums to Aggregate Local Point-Wise Features for Amodal 3D Object Detection

TL;DR

This paper introduces Frustum ConvNet, a novel end-to-end method for amodal 3D object detection from point clouds that leverages 2D proposals to generate frustums, aggregate local features, and accurately estimate 3D bounding boxes.

Contribution

The paper presents a new frustum-based convolutional network architecture for 3D detection that is dataset-agnostic and outperforms existing methods on SUN-RGBD and KITTI datasets.

Findings

Outperforms all existing methods on SUN-RGBD.

Achieves state-of-the-art results on KITTI benchmark.

Demonstrates effectiveness of frustum-based feature aggregation.

Abstract

In this work, we propose a novel method termed \emph{Frustum ConvNet (F-ConvNet)} for amodal 3D object detection from point clouds. Given 2D region proposals in an RGB image, our method first generates a sequence of frustums for each region proposal, and uses the obtained frustums to group local points. F-ConvNet aggregates point-wise features as frustum-level feature vectors, and arrays these feature vectors as a feature map for use of its subsequent component of fully convolutional network (FCN), which spatially fuses frustum-level features and supports an end-to-end and continuous estimation of oriented boxes in the 3D space. We also propose component variants of F-ConvNet, including an FCN variant that extracts multi-resolution frustum features, and a refined use of F-ConvNet over a reduced 3D space. Careful ablation studies verify the efficacy of these component variants. F-ConvNet assumes no prior knowledge of the working 3D environment and is thus dataset-agnostic. We present experiments on both the indoor SUN-RGBD and outdoor KITTI datasets. F-ConvNet outperforms all existing methods on SUN-RGBD, and at the time of submission it outperforms all published works on the KITTI benchmark. Code has been made available at: {\url{https://github.com/zhixinwang/frustum-convnet}.}