On Road Object Importance Estimation A New Dataset and A Model with Multi-Fold Top-Down Guidance

Zhixiong Nan, Yilong Chen, Tianfei Zhou, Tao Xiang

If you like our project, please give us a star ⭐ on GitHub for the latest update.

Multi-Fold Top-Down Guidance Estimation Model

This is the official implementation of the paper "On Road Object Importance Estimation A New Dataset and A Model with Multi-Fold Top-Down Guidance".

This is the quantitative comparison results on TOI and Ohn-Bar datasets.

Update

[2024/9] TOI has been accepted at NeurIPS 2024 as a poster!

Traffic Object Importance (TOI) Dataset

Folder Description:

The object importance annotation results of TOI are located in the "annotation" folder. There are a total of 28 txt files, which correspond to the annotations of 28 videos. The filenames correspond to the raw data in the KITTI official dataset.

To Do:

To use these annotations, you need to download the Raw Data from the KITTI. The files that need to be downloaded are [synced+rectified data].

Citing TOI

If you find our work helpful for your research, please consider citing the following BibTeX entry.

@inproceedings{nanroad,
  title={On-Road Object Importance Estimation: A New Dataset and A Model with Multi-Fold Top-Down Guidance},
  author={Nan, Zhixiong and Chen, Yilong and Zhou, Tianfei and Xiang, Tao},
  booktitle={The Thirty-eighth Annual Conference on Neural Information Processing Systems}
}