Face Mask Detection

The Model is designed to check if any human is wearing a mask or not.

Dataset Description

The Dataset contains a total of 11,792 images divided in three folders

• Test (992 images),
• Train (10,000 images),
• Valid (800 images),

Each of these folders has sub-folders :

• Test

  • WithMask
  • WithoutMask

• Train

  • WithMask
  • WithoutMask

• Valid

  • WithMask
  • WithoutMask

you can download the dataset here

Convolution Base (VGG19)

to import the model, run

  from tensorflow.keras.applications import VGG19

Model Parameters

The total number of parameters of the model are 22,122,049

Result

The training stablized due to early stopping after achieving the following parametrs

The model trained for 16 epochs before early stopping conditions were met.