deepinx / megaface-evaluation

MegaFace Evaluation Tool for MXNet Models

MegaFace is a set of public face datasets published and maintained by the University of Washington's Computer Science and Engineering Laboratory. It is a benchmark of 1 million-scale facial recognition algorithm. The dataset contains 690,572 images with 1,027,060 images. This repository contains a simple tool to evaluate your models on Megaface Challenge 1 benchmark. All the codes are implemented in Python and Mxnet in this repo.

Environment

This repository has been tested under the following environment:

• Python 2.7
• Ubuntu 18.04
• Mxnet-cu90 (==1.3.0)
• OpenCV 2.4.9

Usage

1. Prepare the environment. OpenCV 2.4 is required by the official devkit, for convenience, you can download it from BaiduCloud or GoogleDrive and unzip to /usr/local/lib/opencv2.4.

2. Clone the repository and put your models to be evaluated in models/.

3. Download megaface testpack data from BaiduCloud and unzip it to data/.

4. Download megaface official devkit from BaiduCloud or GoogleDrive and unzip to devkit/.

5. Edit and call run.sh to evaluate your face recognition model performance.

6. Edit and run plot_result.py if you want to visualize the final results.

License

MIT LICENSE

Reference

@inproceedings{kemelmacher2016megaface,
title={The megaface benchmark: 1 million faces for recognition at scale},
author={Kemelmacher-Shlizerman, Ira and Seitz, Steven M and Miller, Daniel and Brossard, Evan},
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
pages={4873--4882},
year={2016}
}

Acknowledgment

The code is adapted based on an intial fork from the insightface repository.