• 

  Training Loss and Accuracy

• 

  Classification Report

• 

  With mask

• 

  Without mask showing alert

COVID FACE MASK DETECTOR.

Inspiration:

Today in this situation face mask is necessary to avoid spread of virus , In my dad company there are 20 workers and to ensure their safety my day asked me to do something and came up with this idea and i have deployed this model in cam using google cloud services and now my dad can able and can ensure that all employee is wearing mask

How I built it :

I built this model with the help of google cloud ,python, tensor flow, MobileNetv2 and I have trained with CNN with 53 layer deep and the model is of high accuracy and easy to use and deploy in large scale.

Accomplishments that I'm proud of :

This is my first Hackathon that I have participated and i have completed my project and made my dad feel proud of me. I have learned to use Tensorflow and deep learning techniques and application on real world problems and I also Satisfied my dad and implement and introduced new technology in our business and I also ensured the safeties' of employee in our company and I am proud of that .

What's next for COVID Face Mask Detector:

My next goal after the hackathon is to expose the dashboard to the user in a way that shows the employee name and shows the picture of the employee and the time that he have not wore mask and it will also alert the employee with the message .

Dependencies :

Python 3.7 Keras 2.2.4 conda install keras=2.2.4 Tensorflow 1.15 conda install tensorflow-gpu=1.15 or conda install tensorflow OpenCV-python ImageDataGEnrator NumPy OS MobileNetV2

Dataset :

I have created the dataset for this project and i have divided it into divided into two categories ie "img_mask" and "img_nomask" and stored into RESORCE folder.

I downloaded the images for both categories from internet and saved them in respective folder in the system. Totally I have 4541 images beacuse we need to get highly accurate model..

img_mask = 2605 img_nomask = 1936

Pre-requisite:

All the dependencies and required libraries are included in the file requirements.txt

Features.

My face mask detector didn't uses any morphed masked images dataset.

The model is accurate, and since we used the MobileNetV2 architecture,

it’s also efficient and thus making it easier to deploy the model

Model accuracy is 0.99

Procedure for usage:

Install Anaconda(Python 3.7) & Tensorflow GPU (I used 1.15 TF)

Then Download this Repository via GIT or zip.

Already i have trained model and uploded in my repository as covid_mask_detector.model

Now copy covid_mask_detector.model and Covid_FaceMAsk_detect.ipynb

Then copy my files to the local disk were jupyer or anaconda navigator is installed

Now run the ipynb file via jupyter it will open your webcam and it will start detection.

For futher Queries contact me at:

vineeth.artifintell@gmail.com

Built With

• jupyter-notebook