✨ Add Landmark Extraction Models

[ishandutta0098]

Problem

We want to support Landmark Extraction for Images and Videos in mukh. To do that the module has been initialized and base class has been written. Refer to mukh/landmarks.

Following two landmark extraction models need to be added -

• BlazeFace
• Mediapipe

Solution

Refer to the Base Class and implement the inference code for landmark extraction for the given models.

Important Note

• Please ask for the issue to be assigned before raising a PR.

Contributing Guidelines

Refer to contributing guide for required contribution instructions.

Share on Socials

If your PR regarding the issue gets merged in mukh feel free to share it on your socials and tag me, I will repost it and help increase your reach especially if you are beginner contributors!

• LinkedIn
• X

[ishandutta0098]

@Vidit-Ostwal shall I assign this to you as discussed?

[Vidit-Ostwal]

@Vidit-Ostwal shall I assign this to you as discussed?

@ishandutta0098 yes, please go ahead

[ishandutta0098]

Cool, I have assigned it to you, lmk if you need any help or discuss and ideas.

[Vidit-Ostwal]

Hi @ishandutta0098, just to confirm the BlazeFace model gives us 17 numbers?

This particular part of the project focuses on the landmark detection, which are the number apart from the first 4.
Let me know If I am missing something.

https://github.com/hollance/BlazeFace-PyTorch
Also if would be great if you could attach some resources in the issue, for anyone new to understand better what exactly the landmark detection means.

[Vidit-Ostwal]

Hi @ishandutta0098, any update on this.