FAQ

The registration form has a checkbox for anonymous/non-anonymous submission. Will we be able to change this at a later date?

Anonymous/non-anonymous status may be changed up until the time of submission. After that point participants must choose what they would like to be.

How many competitors are needed for the competition to proceed to publication?

There need to be at least three competitors to proceed to publication and at least two competitors in a category to declare a winner.

How many times can we submit? only once or many times?

Only once by 30th April 2021

Can we get feedback after we submit? If yes, how can we get it?

Yes of course you would receive feedback after our evaluation of the submitted files. We will send you an email with the feedback.

Besides the RGB image, does the test dataset have deep or infrared images?

The test dataset only has RGB images.

To submit the algorithm, can we use only Python 3.9? Is Python 3.6, 3.7, or 3.8 okay?

Yes, we will accept Python 3.6 or newer releases.

Can you provide a requirement list that contains the needed tools or packages, such as python, torch, numpy, and so on?

Yes, please provide all the necessary information with your submission, which are necessary and would help us run the software solutions at our end.

Are the pictures used in the competition directly cropped? Or provide the key points of the face?

No they are not cropped and no key points of the face are provided.

Will you consider the models that use background information rather than facial features?

We have no reservations about how your software solution detects attacks. It is really up to the participants.

How can we provide an executable python file for Linux? Do you have any documentation that you can recommend? Would pyinstaller be enough to create this executable?

We do not have any guidelines for executable submission. We will accept submissions for Windows, Linux, and macOS. You may submit a python script as well with instructions to run the script.

If you have questions feel free to email us directly at datasets@livdet.org or send your question via the contact us page.