Facial Mask recognition

Its a project in which using Haar-cascade frontal face algorithm the face is recoginzed, processed and marked as masked or un-masked in the given video uploaded or through a Web Cam.

Comment

Inspiration

Basically, COVID-19 was my inspiration in the sense the sheer amount of people who died during the pandemic is due to people's negligence this could have been avoided if an AI models were used in public places to recognize the negligent people who weren't wearing mask and take strict action against them.

What it does

The face of a person is recognized and marked in the given video uploaded or captured via web cam and tells whether the person is wearing mask or not.

How i built it

Basically i built the project in 2 steps

  • Step 1: Creating and training the model to recognize the masked and unmasked faces using given data sets.
  • Step 2: Implementing the model and giving out the desired output based on predictions of the model.

Challenges i ran into

  • The time required to train my model was very long.
  • To increase accuracy of the model.

Accomplishments that we're proud of

  • test accuracy : 91.00% test loss: 0.16

What we learned

  • i learnt about Sequential model, harcasscade-frontal face algorithm , numpy , tensorflow , image processing and OpenCV.

What's next for Facial Mask recognition

  • To upgrade my project, basically had an idea to integrate the model with CCTV cameras and use a warning system to alert the hosts.
  • To improve the accuracy of my model.

Built With

Share this project:

Updates