Real Time Gesture Detection and Recognition using Tensorflow.js

Traditional methods of inputting data to systems or other digital gadgets to receive the desired outputs are no longer cutting-edge and progressive. Usable command set gets limited because of the limited usage of input devices such as a mouse. One of the solutions to these drawbacks is usage of gestures for inputting data and commands to the digital gadgets. In this paper we have presented an approach via which the devices become capable of detecting or recognizing the gestures and hence are able to perform the operations or tasks as desired by the users. Even in the surveillance systems, the gestures of humans are cogitated to be highly applicable information. In this paper we have discussed about gestures, it forms, its advantages and applications in today’s world. We have made a model for the same using TensorFlow.JS Web library and have implemented Posenet algorithms. Anticipated method is implemented in Atom software.