American Sign Language recognition using convolutional neural networks.
Make the run.sh script executable.
chmod u+x run.sh
The run script assumes python 3.x is being used and it is accessible as python. If not, change all instances of python to python3 in the run script.
To train the model:
./run.sh train
To evaluate the model:
./run.sh evaluate
To predict on real time data:
./run.sh predict
Note: You will need to install the IP Webcam link app or other similar application on your android smartphone to access the camera for images. Or modify the predict script according to your needs.
The original data was obtained from Kaggle. We have included out modified data in the repository.
Uncompress it into the data directory.
The directory structure should look like this:
root
├── Code
│ ├── evaluate.py
│ ├── model
│ │ └──
│ ├── modelling.py
│ ├── predict.py
│ ├── preprocess.py
│ └── train.py
├── data
│ └── augmented.csv
├── Group-Proposal
│ └── Final Project Proposal.pdf
├── ishita-jain-individual-project
│ ├── Code
│ │ └── train1-ij.py
│ └── Individual-Final-Project-Report
│ └── ishita-jain-final-report.pdf
├── Jyothi-Ganji-individual-project
│ ├── Code
│ │ └── mywork.py
│ └── Individual-Final-Project-Report
│ └── Jyothi-Ganji-final-project.pdf
├── Final-Group-Presentation
│ └── final presentation.pdf
├── Final-Group-Project-Report
│ └── FINAL REPORT.pdf
├── Mihir-Gadgil-individual-project
│ └── Individual-Final-Project-Report
│ └── Mihir-Gadgil-final-project.pdf
├── README.md
└── run.sh