WRITE IN AIR

This project proposes an effective approach for writing in air using just the index fingertip. Our system uses an ordinary computer webcam to track the fingertip trajectory. A CNN model does object detection to classify the trajectory into 12 classes, 10 digits and 2 mathematical operators for addition and subtraction. Finally, these predictions are used for mathematical expression evaluation.

About the files

1. src : This directory contains source code

• main.py : Contains main program for finger detection to expression evaluation
• predict.py : Contains driver code for predictions of trajectory using trained model
• CNNModel.ipynb : Code for training the CNN model

2 .weights : This directory contains trained weigths and model

• final_weights.h5 : Contains trained weigths of the model
• final_model_weight.h5 : Contains trained weights and model

3. data.zip : Dataset

4. outputs : This folder contains the various images of digits generated from the trajectory.

5. requirements.txt : This file contains all the requirements.

How To Use

On Ubuntu

1. Clone the repository from github

2. Create a Python virtual environment

$ python3 -m venv Write-In-Air
$ cd Write-In-Air
$ source bin/activate

3. Install requirements by using requirement.txt

$ pip3 install -r requirements.txt 

4. Run the program

$ python3 main.py

4. Keep your hand in the given rectangle and start writing

On Windows

Install the requirements provided in requirements.txt in Anaconda and run the program.

### Note: For this code to work properly, clear backgroud and bright lighting is necessary.