# ImVoxelNet: Image to Voxels Projection for Monocular and Multi-View General-Purpose 3D Object Detection

## Introduction

<!-- [ALGORITHM] -->

We implement a monocular 3D detector ImVoxelNet and provide its results and checkpoints on KITTI dataset.
Results for SUN RGB-D, ScanNet and nuScenes are currently available in ImVoxelNet authors
[repo](https://github.com/saic-vul/imvoxelnet) (based on mmdetection3d).

```
@inproceedings{rukhovich2022imvoxelnet,
  title={Imvoxelnet: Image to voxels projection for monocular and multi-view general-purpose 3d object detection},
  author={Rukhovich, Danila and Vorontsova, Anna and Konushin, Anton},
  booktitle={Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision},
  pages={2397--2406},
  year={2022}
}
```

## Results

### KITTI

|  Backbone   |Class| Lr schd | Mem (GB) | Inf time (fps) | mAP | Download |
| :---------: | :-----: |:-----: | :------: | :------------: | :----: |:----: |
| [ResNet-50](./imvoxelnet_kitti-3d-car.py) | Car | 3x | | |17.4|[model](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/imvoxelnet/imvoxelnet_kitti-3d-car_20210610_152323-b9abba85.pth) &#124; [log](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/imvoxelnet/imvoxelnet_kitti-3d-car_20210610_152323.log.json)|