TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK	REMOVE
3D Semantic Scene Completion	KITTI-360	TPVFormer	mIoU	13.64	# 5
Prediction Of Occupancy Grid Maps	nuScenes	TPVFormer04	mIoU	52.058	# 1

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/tri-perspective-view-for-vision-based-3d/prediction-of-occupancy-grid-maps-on-nuscenes)](https://paperswithcode.com/sota/prediction-of-occupancy-grid-maps-on-nuscenes?p=tri-perspective-view-for-vision-based-3d)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/tri-perspective-view-for-vision-based-3d/3d-semantic-scene-completion-on-kitti-360)](https://paperswithcode.com/sota/3d-semantic-scene-completion-on-kitti-360?p=tri-perspective-view-for-vision-based-3d)`

Tri-Perspective View for Vision-Based 3D Semantic Occupancy Prediction

CVPR 2023 · Yuanhui Huang, Wenzhao Zheng, Yunpeng Zhang, Jie zhou, Jiwen Lu ·

Modern methods for vision-centric autonomous driving perception widely adopt the bird's-eye-view (BEV) representation to describe a 3D scene. Despite its better efficiency than voxel representation, it has difficulty describing the fine-grained 3D structure of a scene with a single plane. To address this, we propose a tri-perspective view (TPV) representation which accompanies BEV with two additional perpendicular planes. We model each point in the 3D space by summing its projected features on the three planes. To lift image features to the 3D TPV space, we further propose a transformer-based TPV encoder (TPVFormer) to obtain the TPV features effectively. We employ the attention mechanism to aggregate the image features corresponding to each query in each TPV plane. Experiments show that our model trained with sparse supervision effectively predicts the semantic occupancy for all voxels. We demonstrate for the first time that using only camera inputs can achieve comparable performance with LiDAR-based methods on the LiDAR segmentation task on nuScenes. Code: https://github.com/wzzheng/TPVFormer.