3M3D: Multi-View, Multi-Path, Multi-Representation for 3D Object Detection

Jongwoo Park, Apoorv Singh, Varun Bankiti

Abstract: 3D visual perception tasks based on multi-camera images are essential for autonomous driving systems. The latest work in this field performs 3D object detection by leveraging multi-view images as an input and iteratively enhancing object queries (object proposals) by cross-attending multi-view features. However, individual backbone features are not updated with multi-view features, and it stays as a mere collection of the output of the single-image backbone network. Therefore we propose 3M3D: A Multi-view, Multi-path, Multi-representation for 3D Object Detection where we update both multi-view features and query features to enhance the representation of the scene in both fine panoramic view and coarse global view. Firstly, we update multi-view features by multi-view axis self-attention. It will incorporate panoramic information in the multi-view features and enhance understanding of the global scene. Secondly, we update multi-view features by self-attention of the Region of Interest (ROI) windows which encodes local finer details in the features. It will help exchange the information not only along the multi-view axis but also along the other spatial dimension. Lastly, we leverage the fact of the multi-representation of queries (MRQ) in different domains to further boost performance. Here we use sparse floating queries along with dense Bird’s Eye View (BEV) queries, which are later post-processed to filter duplicate detections. Moreover, we show performance improvements on the nuScenes benchmark dataset [1] on top of our baselines.