Skip to content

Releases: PovertyAction/Facial-Validation

Version 0.1 for Windows

09 Feb 15:55
@fhalamos fhalamos
v0.1
c71a70d
Compare
Choose a tag to compare
Version 0.1 for Windows Latest
Latest 
v0.1

Adapting for gui_less app to work without survey.dta
Assets 3
Loading

Windows Initial Alpha Release

15 Feb 19:33
@joshjacobson joshjacobson
v0.1.0
fd2d028
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
Windows Initial Alpha Release Pre-release
Pre-release

Facial Validation

About

This application detects whether listings of paired pictures are of the same person. It is most often used to help ensure the same person is being interviewed across waves or to detect when someone enrolls more than once in a program. Though this tool can be helpful, ensuring identity is ultimately still your responsibility.

As an alpha release, it is expected that this will contain bugs and will not work for all users. Please share any issues or feedback you have resulting from use by filing an issue on GitHub or emailing. Please do not share this application outside of IPA at this time.

This tool is current listed as an alpha release because it is completely untested on IPA field pictures due to a lack of data access. You can help make this tool better, or have it customized to your needs, by working with GRDS on your upcoming project that involves participant photos.

How to Use

  1. Take photographs of participants in line with the guidelines.
  • The best way to ensure success is to train enumerators, pilot field photography and data collection, and work with GRDS to ensure the pictures collected during pilot are able to be successfully analyzed by the tool. GRDS can also continue to monitor photo quality throughout data collection.

  1. Place all images in a single folder on your machine and prepare the input template. You'll provide the path to the directory containing the images, and paired image filenames to compare.

  2. Download the .exe application file from the latest release. Run the application and follow the instructions to upload the input template. It may take a little while (up to a minute) for the main application to load and open.

  3. Examine the output file, results.csv in the folder containing the images. A '1' on the threshold test denotes that the person in the image pair is the same, while a '0' indicates that they appear not to be.

Help and Support

Please check the issues section on GitHub for previously solved issues. If no solution is available there feel free to open an issue; the author will attempt to respond in a reasonably timely fashion. If your request is urgent you may contact [email protected]

Contributing

We welcome contributions in any form! To contribute please fork the project make your changes and submit a pull request. We will do our best to work through any issues with you and get your code merged into the main branch.

Citing

If you have used this application in a study or publication, please use the following:

Jacobson, J. Innovations for Poverty Action: Facial Validation. 2017. https://github.com/PovertyAction/PII_detection

Credit

King, D. High Quality Face Recognition with Deep Metric Learning. [online] Blog.dlib.net. Available at: http://blog.dlib.net/2017/02/high-quality-face-recognition-with-deep.html [Accessed 8 Feb. 2018].

Licensing

This project is MIT Licensed.

Assets 3
Loading