Omani Research Project Develops Hand Gesture Recognition System

Muscat, A research team from Sohar University developed a hand gesture recognition system based on neural networks, image processing techniques, and Internet of Things (IoT). The system detects hand gestures and transfers the gestures to a written text then to audio.

The research team consists Dr. Abdallah Abualkishik, Sara Salim Al Khaifi and Mojtaba Jaffar.

Dr. Abdallah Abualkishik said that the research team developed a recognition system that utilized the convolutional neural network method. It aims to convert the dynamic deaf signs from live video to text and speech using a Raspberry Pi device and a normal camera. The researchers created the dataset for this project. It contains 62000 (64×64-pixel) images of the 30 letters of the alphabet and 1 word. Each pattern (Letter) has 2000 images that are divided into 1750 images for training and 250 images for testing.

Dr. Abdallah Abualkishik also pointed out that the research was funded by the Ministry of Higher Education, Research and Innovation.

Source: Oman News Agency