CenterAtt: Fast 2-stage Center Attention Network

TL;DR

This paper presents CenterAtt, a fast 2-stage center attention network for real-time 3D detection, built upon CenterPoint, with various optimizations to enhance speed and efficiency on the Waymo dataset.

Contribution

Introduces CenterAtt with center attention head and feature pyramid network, optimizing for real-time detection using techniques like batchnorm merge and GPU-accelerated voxelization.

Findings

Achieved 6th place in real-time 3D detection challenge

Implemented effective speed-up techniques for 3D detection

Enhanced CenterPoint framework with center attention modules

Abstract

In this technical report, we introduce the methods of HIKVISION_LiDAR_Det in the challenge of waymo open dataset real-time 3D detection. Our solution for the competition are built upon Centerpoint 3D detection framework. Several variants of CenterPoint are explored, including center attention head and feature pyramid network neck. In order to achieve real time detection, methods like batchnorm merge, half-precision floating point network and GPU-accelerated voxelization process are adopted. By using these methods, our team ranks 6th among all the methods on real-time 3D detection challenge in the waymo open dataset.