American Sign Language (ASL) Detection utilizes a Convolutional Neural Network (CNN) model for accurate interpretation. By analyzing video input, the CNN identifies and classifies hand gestures corresponding to ASL signs. This technology enables real-time translation of ASL into written or spoken language, enhancing communication accessibility for the hearing impaired and fostering inclusivity in various domains such as education, healthcare, and social interaction.
The dataset used can be found on Kaggle
-
├───input │ ├───test │ │ └───A │ | └───B │ | ... │ ├───train │ │ └───A │ │ └───B | | ... │ │ │ └───preprocessed_image │ ├───A │ ├───B │ ... ├───outputs │ ├───model.pth │ ├───loss_plot.png │ ├───accuracy_plot.png │ ├───A1_Test.jpg │ ├───B1_Test.jpg | ... └───src │ test-with-image.py │ test-with-camera.py │ cnn_models.py │ create_csv.py │ preprocess_image.py │ train.py
