Smart Feature is What You Need

TL;DR

This paper introduces MMA, a novel point cloud feature network that enhances weakly-supervised 3D object detection by leveraging shape and object inference, significantly improving performance and robustness.

Contribution

The paper proposes MMA, a new feature representation network that combines adjacency and disparity attention to improve weakly-supervised 3D detection accuracy.

Findings

MMA improves weakly-supervised detection to approach fully-supervised performance.

MMA effectively mitigates label jitter, turning it into a data enhancement source.

Extensive experiments validate MMA's effectiveness in indoor scenarios.

Abstract

Lack of shape guidance and label jitter caused by information deficiency of weak label are the main problems in 3D weakly-supervised object detection. Current weakly-supervised models often use heuristics or assumptions methods to infer information from weak labels without taking advantage of the inherent clues of weakly-supervised and fully-supervised methods, thus it is difficult to explore a method that combines data utilization efficiency and model accuracy. In an attempt to address these issues, we propose a novel plug-and-in point cloud feature representation network called Multi-scale Mixed Attention (MMA). MMA utilizes adjacency attention within neighborhoods and disparity attention at different density scales to build a feature representation network. The smart feature representation obtained from MMA has shape tendency and object existence area inference, which can constrain the region of the detection boxes, thereby alleviating the problems caused by the information default of weak labels. Extensive experiments show that in indoor weak label scenarios, the fully-supervised network can perform close to that of the weakly-supervised network merely through the improvement of point feature by MMA. At the same time, MMA can turn waste into treasure, reversing the label jitter problem that originally interfered with weakly-supervised detection into the source of data enhancement, strengthening the performance of existing weak supervision detection methods. Our code is available at https://github.com/hzx-9894/MMA.