Real Time Sign Language Detection Using ESP32-CAM

This project focuses on real-time sign language detection, utilizing an ESP32-CAM board for video capturing.

Two methods are implemented:

1. NickNochnack’s Method: A custom dataset of 100 images (20 images per class) is created with five classes: “hello,” “welcome,” “to,” “my,” and “website.” I wrote a separate code for capturing and gathering this dataset using laptop’s camera. This dataset is then used to train NickNochnack’s sign language detection model.

2. Daniel’s Tutorial: This method also uses a different custom dataset captured and gathered using a separate code. It focuses on detecting hands and fingers, followed by training the model to recognize numerical signs (0-9).

This project demonstrates a hardware-software integration for real-time gesture recognition and expands accessibility through sign language detection. The code and additional details can be found in the GitHub repository.

Note 1: I wrote code for capturing the dataset and used my laptop for gathering my dataset. It is possible to use your phone, directly use the ESP32-CAM, or any other method.

Note 2: For annotation, use Pascal VOC format. Otherwise, you need to modify a couple of lines in the code.

Link to the gihub repository:

Link