Awesome

Size-Invariant Metrics

This is the official code for the computation of Size-Invariant Metrics in paper "Size-invariance Matters: Rethinking Metrics and Losses for Imbalanced Multi-object Salient Object Detection" accepted by International Conference on Machine Learning (ICML2024). The paper is available here, and the complete repository is here.

With this repository, SI Metrics can be directly computed given the path to the prediction and ground truth directory.

Datasets

All that you need to prepare is a directory containing the ground mask "xxx.png", and a directory containing the prediction mask "xxx.png".

Here some notifications:

1. The range of prediction map can be either in [0, 255] or in [0, 1]. If the range is [0, 255], set the normalize to True in config.yaml.
2. The epsilon in config.yaml is designed to remove small noise points in the ground truth map. If there is no connected component found after the denoising, just set epsilon to a larger value.

Evaluation

The evaluation configs are stored at config.yaml, where you can modify the settings of data, model, visualization, and metrics.

To begin evaluation, you can run the following command:

python main.py

Citation

If you find this work or repository useful, please cite the following:

@inproceedings{li2024sizeinvariance,
title={Size-invariance Matters: Rethinking Metrics and Losses for Imbalanced Multi-object Salient Object Detection}, 
author={Feiran Li and Qianqian Xu and Shilong Bao and Zhiyong Yang and Runmin Cong and Xiaochun Cao and Qingming Huang},booktitle={The Forty-first International Conference on Machine Learning},
year={2024}
}

Contact us

If you have any detailed questions or suggestions, feel free to email us: lifeiran@iie.ac.cn! Thanks for your interest in our work!