GestureViz: A Mobile Application for Vision-based Hand Gesture Recognition


    Inventor

    “The aged population in Malaysia is expected to rise at the quickest rate in the future decades due to historical and current trends in fertility and mortality.” According to the Department of Statistics Malaysia, we will have an ageing population of 5.6 million seniors, or 15 percent of the total population, by 2035. In 2023, it is estimated that the elderly population in Malaysia will constitute 8.5% of the total population, equivalent to 290,000 people. The increasing aging population has also increased the need for eldercare services and facilities. It is thus important to create an inclusive and age-friendly smart living environments. AI enables innovative solutions to support the elderly population within cities and communities. AI-powered systems can assist in healthcare monitoring, fall detection, medication management, and providing companionship to seniors, contributing to their well-being and independence. This project targets to assist seniors with mobility impairment. Due to this limitation, it makes it difficult for them to request help when they are in need from afar. While the elderly’s caretaker is responsible for assisting them with their daily tasks; however, it would be difficult for them to attend to the elderly at all times. In this situation, a hand gesture system can solve this communication barrier. Seniors can perform specific hand gestures that represent their requests, and the recognition system will detect these inputs and send an alert to the caretaker's mobile device. This technology can also facilitate caretakers to fulfil their responsibilities in a more efficient manner.Our solution falls into the people-centric technology category because it is designed and personalized to meet the needs and preferences particularly of the elderly and their caretakers. Our proposal is an AI-assisted solution constituting of an ubiquitous real-time alert system that captures hand gestures from a camera of a single board computer. The innovative component is that our solution is privacy preserving, that is the live video captured via the board is not shared and not kept neither saved on clouds. Instead the cloud server only transfers the interpretation of the action or meaning of the hand gesture. The interpretation module is located and processed within the single board computer. By using hand gesture, it lets user to have a communication that is more natural and intuitive, without the need for physical buttons or switches. This can be particularly helpful for seniors with mobility impairments or difficulty with fine motor control. Additionally, a hand gesture system can be customized to a user's specific gestures, making it more responsive to their individual needs and preferences. This level of personalization can be especially valuable for elderly individuals who have difficulty using traditional input methods. A feedback mechanism is incorporated with the system in such a way that the gesture action being captured is displayed and properly labelled on the user's side via the display device (if needed). On the caretaker's side of the system, the mobile application only receives interpreted gesture that can run on multi-platform (Android, iOS) and multi-device (mobile, tablet, PC). The gesture interpretation module operates a machine learning back-end, which is reliable and gives high accuracy, average of close to 98%. Furthermore, it can constantly be developed thus making personalization and future enhancements possible. We can see that the solution can be implemented at elderly nursing homes and has the potential to be expanded to mass population as it uses AI-assisted technologies via mobile devices. Assuming that 5% of the elderly nursing homes population requires immobility elderly care, it is expected that this project will contribute to around 29,000 elderly in Malaysia.
  • Description of Invention

    Dr. Wan Noorshahida Mohd Isa & Dr Noramiza Hashim

  • Intellectual Property (IP) Status

    • TRL Status: 4

Video