DeepInteraction: 3D Object Detection via Modality Interaction

TL;DR

DeepInteraction introduces a novel modality interaction strategy for 3D object detection that maintains individual modality representations, leading to significant performance improvements on the nuScenes dataset.

Contribution

The paper proposes a new modality interaction approach with a dedicated architecture to better exploit modality-specific information in 3D detection.

Findings

Outperforms prior methods on nuScenes dataset

Achieves first place on nuScenes leaderboard

Demonstrates significant accuracy improvements

Abstract

Existing top-performance 3D object detectors typically rely on the multi-modal fusion strategy. This design is however fundamentally restricted due to overlooking the modality-specific useful information and finally hampering the model performance. To address this limitation, in this work we introduce a novel modality interaction strategy where individual per-modality representations are learned and maintained throughout for enabling their unique characteristics to be exploited during object detection. To realize this proposed strategy, we design a DeepInteraction architecture characterized by a multi-modal representational interaction encoder and a multi-modal predictive interaction decoder. Experiments on the large-scale nuScenes dataset show that our proposed method surpasses all prior arts often by a large margin. Crucially, our method is ranked at the first position at the highly competitive nuScenes object detection leaderboard.