Blinking to a New Era of Environmental Interaction Project-Bitrac

This paper introduces an innovative solution for a person to interact with his surroundings with the power of blink. We utilize various modern algorithms and models like Inception v3, a pre-trained Deep Neural Network model for image classification [5], and facial landmark detection algorithm of Kazemi and Sullivan (2014) trained using various datasets like the IBUG dataset to efficiently capture an activation signal in the form of a blink and thereby capture the image of the object to be interacted with. This is achieved using two micro cameras, one on a watch and one hooked to the user. The image is then processed, and the appliance is thereby switched on with an IR signal. A disabled person can deeply interact with all the appliances with the help of a few blink patterns.