search

thecodermaniac / FaceApi

4 stars 5 forks source link

face detection along with object detection preferably mobile, tablet etc #9

Open thecodermaniac opened 11 months ago

thecodermaniac commented 11 months ago

Functionality details

  1. add an api endpoint which is able to detect human faces preferably with mobile, tablet in front of screen
  2. May use tensorflow ml models for the task or import other 3rd party models
  3. need to be fast as there will be real time detection

Usage

  1. generally will be used for automatic exam proctoring.

How it will be used

  1. The frontend sends an image from the web cam at intervals and it needs to be verified.
  2. All things will be happening in real time (you are required not to implement the front-end part)

Reference:- object detection

ASHUdev05 commented 11 months ago

Can I work on this as I previously explored a similar issue (:

thecodermaniac commented 11 months ago

Can I work on this as I previously explored a similar issue (:

yeah sure assigned

ASHUdev05 commented 11 months ago

I am wondering how can we fetch the facial data for training? Certain datasets are available like https://github.com/microsoft/DigiFace1M , but such a huge data would not be required for this usecase atm i guess, moreover we do not have such hardware capabilities to train the model on such an extensive dataset.

thecodermaniac commented 11 months ago

I am wondering how can we fetch the facial data for training? Certain datasets are available like https://github.com/microsoft/DigiFace1M , but such a huge data would not be required for this usecase atm i guess, moreover we do not have such hardware capabilities to train the model on such an extensive dataset.

you don't need to train any model its time taking and hardware heavy

Few suggestion to do the task

  1. use two model one for face and other for mobile or table and same image is passed to the both models
  2. use yolo model its fast and does both prediction at the same time.