Customizable Dynamic Hand Gesture recognition System for Motor Impaired people using Siamese neural network

Every year, between 250000 and 500000 people across the world suffer a major spinal cord injury, many of which result in motor impairment. A set of physical gestures can comprise an entire language which makes it a powerful form of communication for people who suffer from motor impairment. Most gesture recognition systems require the user to learn specific gestures prescribed by the system. This turns out to be a major disadvantage of these systems as most motor impaired individuals are heavily constrained in terms of movement. Hence they may not be able to accomplish certain gestures defined by the system. So in this paper we propose a realtime system which can be customized to user-specific dynamic hand movements. This system can be used to carry out certain tasks. Haar Cascade detection algorithm was used in order to track the movements of the hand and trace the path. Siamese neural network was used for the purpose of customization and recognition of the gestures.