The project aims at building a machine learning model that will be able to classify the various hand gestures used for fingerspelling in sign language. In this user independent model, classification machine learning algorithms are trained using a set of image data and testing is done on a completely different set of data.
For the image dataset, depth images are used, which gave better results than some of the previous literatures, owing to the reduced pre-processing time.
Various machine learning algorithms are applied on the datasets, including Convolutional Neural Network (CNN). An attempt is made to increase the accuracy of the CNN model by pretraining it on the dataset. However, a small dataset was used for pre-training, which gave an accuracy of 85% during training.
In this project we had predicated American Sign Language (ASL). In this project we have predicated numbers[0-9], letters[A-Y] and 10 basic gestures ["Don't Want","Hello","Help","I Love You","Money","No","Ok", "Thank You","Want","Yes"].
We use openCV to get input through webcam. We detect landmark through cvzone library. After detecting landmark, once we save the gesture in folder it gives the meaning to the gesture. 250 to 300 pictures are given to individual gesture for accuracy. We have taken each picture in teachable machine or training machine and trained it and converted it into one model file respectively and gave its path in the output part of code. Lastly through openCV we access webcam and then the user shows the sign which is predicted and analyzed to give the final detected result.
This our Sem-3 mini project.
-
OS: Windows 10 and 11
-
Processor: i5 and 8th generation
-
RAM 8GB
-
python 3.10
-
Tensorflow framework- 2.9.1,keras API -2.9.0
-
Real-time Computer vision using openCV - 4.6
-
Google teachable Machine for Training part
-
Numpy as np
-
cvzone - 1.5.6
-
Hands detector using matplotlib - 3.5.3
- Create Dataset (In this project I have made my own dataset)
- Train the dataset
- Predicate the sign
| S.No. | Name | Role | GitHub Username |
|---|---|---|---|
| 1. | Neha Singh | Full Stack Web Developer | @devgeek2700 |
| 2. | Shalaka Kadam | Research and Design | @shalaka2603 |
| 3. | Amulya Shetty | Product Designer | @AmulyaShetty11 |
Python, openCV....
Contributions to the Healthcare Consultant App are welcome! If you'd like to contribute, please fork the repository, make your changes, and submit a pull request. Be sure to follow the project's coding standards and guidelines.
This project is licensed under the MIT License, which means you are free to use, modify, and distribute the code for personal or commercial purposes. See the LICENSE file for more details.