This repository aggregates academic papers on bird’s-eye-view (BEV) perception for autonomous vehicles, covering multi-camera 3D object detection and segmentation models. It includes implementations of transformer-based approaches like BEVFormer and DETR3D, as well as depth estimation methods like BEVDepth. The papers span venues including CVPR, ICCV, and ECCV, representing state-of-the-art computer vision research for perception in self-driving systems.