Augmented reality for industrial processes automation

In this project I was responsible for developing deep learning algorithms for speech enhancement, keyword spotting and gesture recognition to be used on AR platform for industrial processes automation. The AR platform will make it possible to increase the productivity of employees working with production machines and facilitate knowledge transfer and reduce the level of training required to perform service work. It can also serve as a training platform for new employees as well as a tool supporting the work of engineers. The complete software platform will work with dedicated AR devices equipped with a 3D space mapping system. The developed system will be supported by algorithms based on deep learning methods for the processing of audiovisual data.

Main tasks:

• Development of speech denoising algorithms based on machine learning methods.
• Development of KWS (Keyword spotting) algorithms
• Development of hand tracking algorithms and gesture recognition.

Technologies used: Python, Pytorch

Methods used: Deep Neural Networks, Keyword Spotting, Speech Denoising, Hand Detection and Gesture Recognition

Learn more about the project

Project’s website: neurosys.com