Social Distancing and Face Mask Detection :mask:

Developed By:

👉  [Shantanu Gupta]    Portfolio       GitHub      LinkedIN

About--

                  Social Distancing and Face Mask Detection Platform utilizes Artificial Network to perceive if a person walk with maintain social distance and does/doesn’t wear a mask as well. The application can be associated with any current or new IP cameras to identify individuals maintaining social distance with/without a mask. :mask:

System Requirement -- :desktop_computer:

• SOFTWARE--      
  • Software:                 Anaconda + Python 3.x (3.8 or earlier)
  • Editor:                      VS Code/ PyCharm/ Sublime/ Spyder
  • Environment:          TensorFlow
  • GPU Drivers:           Nvidia® CUDA® 11.0 requires 450.x or above                                CUDA® Toolkit (TensorFlow >= 2.4.0)                                cuDNN SDK 8.0.4 (TensorFlow >= 2.4.0)

• HARDWARE--
  • GPU:                                     Graphics Processor (NVIDIA) ̶min 2GB
  • Camera:                                CCTV/ Webcam/ Mobile Camera (Sharing Camera)
  • Storage Disk (Optional):       SSD – Min 400MB/s Read Speed

Installation Process--

• Download Anaconda Software --

:loudspeaker: During Installation be sure to check to set Environmental variable path

• Create new Environment for the installation of libraries:
  • Open Command Prompt / Anaconda Prompt and type conda create --name tf_python
    you can set any name in place of tf_python to create a new envionment. and after type y and enter.
  • Install all required Libraries given in requirement.txt by using command pip install -r requirement.txt

Required Libraries--

•          TensorFlow
•    TensorFlow-GPU
•                  Python
•        Scikit Learn
•         Computer Vision
•                  Scientific Python

File Required to Download --

• DATASETS :         Using datasets to train the model for Face Mask Detection model. To download the dataset -- :point_right: Click here  :point_left: (Dataset with 4,000 Images Sampels) :star2:File contain 2 Sub-Folder i.e. With_mask & Without_mask (each folder contain 2k samples of images).

This is a balanced dataset containing faces with and without masks with a mean height of 283.68 and mean width of 278.77

• Yolo Weights (V3) -- Pre-Trained model:
          YOLO (You Only Live Once), the pre-trained weights of the neural network are stored in yolov3.weights                                     Download the Weight File :point_right: Click here :point_left:

Trained Result of Face Mask Model--

File Structure

Set all downloaded files to their respective folders/path as given in Folder Structure Diagram.

RUN the Main Module--

• Using Command Prompt or Anaconda Prompt:
  • To activate environment:--      conda activate tf_python
  • Run main module:-- python main.py

Outputs--

Contribute:

                :fire: Contributions are always welcome!

Drop a :star: if you like this Repository.. :smile:

:point_right::point_right::point_right:       []:point_left::point_left::point_left: