Submission

All algorithm submissions will be done via email.

  1. Please zip your trained algorithm in a folder. Name of the folder should be your full name.
  2. Upload the file to a cloud storage. Share the downloadable link with us via email at datasets@livdet.org.
  3. Please share any specific details to download your file if it is encrypted or requires specific knowledge.

Each Algorithm should have the following parameters:

LIVENESS [input file] [outputfile]

LIVENESS [input file] [outputfile]

LIVENESS is the executable name.

Format: Windows, Ubuntu, Mac OS

[inputfile]

  • For the single image category, the [input file] will be a txt file with the list of images to analyze. Each image will be in either .jpg or .png format.
  • For video category, the [input file] will be a txt file with a list of videos to analyze. Each video will be up to a five seconds in .mp4 format (up to 60 fps).

[outputfile]

  • Txt file with the output of each processed image/video with a “return” between each output, in the same order as the input file.
  • There will be one output file for each input file.
  • The output is an integer value representing a posterior probability of the live class given in the image or a degree of “liveness” normalized in the range 0 to 100. With 100 being the maximum degree of liveness and 0 meaning the image is fake and a threshold being set at 50.
  • In the case that the algorithm has not been able to process the image/video, the correspondent output must be -1000 (failure to enroll).
  • Each user can configure their algorithm by using public datasets.
  • Participants may also choose to publish a description and/or source code of their algorithm, but this is not mandatory.

If you are unable to submit an executable, we will accept Python (v. 3.6 or newer) and MATLAB (2020a or above) codes.

If you have any queries regarding algorithm submission, please feel free to reach out at datasets@livdet.org