Whisper Hands

The "Whisper Hands" project aims to develop an innovative wearable device that recognizes Palestinian Sign Language (PSL) gestures in real time, facilitating communication for individuals with hearing and speech impairments. The system utilizes an ESP32-WROOM microcontroller embedded in a glove worn on the dominant hand, with flex sensors placed on key fingers to capture gestures. The collected data is used to train a gesture recognition model using Edge Impulse Studio, allowing the device to process gestures without an internet connection. "Whisper Hands" introduces a wearable solution that employs embedded AI to recognize Palestinian Sign Language in real time, advancing inclusive technology for sign language interpretation. A distinguishing feature of our project is the use of the Conv1 algorithm, which enhances the accuracy and efficiency of gesture recognition. This project enhances human-computer interaction (HCI) and assistive technology by effectively deploying machine learning on edge devices for sign language recognition. We provide a new dataset and model specifically for PSL, contributing valuable insights to the fields of sign language processing and wearable AI. This project addresses communication barriers for the deaf and hard-of-hearing community in Palestine, offering an affordable and localized technology. Initial results show high accuracy in recognizing PSL gestures with low latency. We aim to create a fully functional prototype that serves as a foundation for scalable assistive technology, with plans to expand gesture coverage and integrate text or speech output for seamless interaction. Given the promising initial results, we plan to begin drafting a research paper to publish the contributions made through the project.