Paint and Distill: Boosting 3D Object Detection with Semantic Passing Network

TL;DR

This paper introduces SPNet, a semantic passing framework that enhances lidar-based 3D object detection by leveraging semantic knowledge from ground-truth labels without additional inference costs, leading to improved accuracy.

Contribution

The novel semantic passing framework guides lidar models using semantic knowledge from labels, boosting detection performance without extra inference overhead.

Findings

SPNet achieves 1-5% AP improvement on KITTI benchmark.

Seamless integration with existing 3D detection frameworks.

State-of-the-art results on KITTI test set.

Abstract

3D object detection task from lidar or camera sensors is essential for autonomous driving. Pioneer attempts at multi-modality fusion complement the sparse lidar point clouds with rich semantic texture information from images at the cost of extra network designs and overhead. In this work, we propose a novel semantic passing framework, named SPNet, to boost the performance of existing lidar-based 3D detection models with the guidance of rich context painting, with no extra computation cost during inference. Our key design is to first exploit the potential instructive semantic knowledge within the ground-truth labels by training a semantic-painted teacher model and then guide the pure-lidar network to learn the semantic-painted representation via knowledge passing modules at different granularities: class-wise passing, pixel-wise passing and instance-wise passing. Experimental results show that the proposed SPNet can seamlessly cooperate with most existing 3D detection frameworks with 1~5% AP gain and even achieve new state-of-the-art 3D detection performance on the KITTI test benchmark. Code is available at: https://github.com/jb892/SPNet.