Smart-spaces

Machine learning from radar and camera to model shared spaces and optimize their utilization. This project was prepared during Junction 2019

Description

Machine learning from radar and camera to model shared spaces and optimize their utilization. Radar data is used to monitor the number of people in a specific area and this information is used to model the space occupancy over time. This model is used to adjust the best global settings of light and room temperature conditions. For personalized or local adjustments, the high-resolution radar is used to estimate the user pose and recognizes gesture commands.

In this prototype we used a webcam as a proxy to radar detection, PoseNet model for user pose extraction, regression model for global settings prediction.

Demo

Team

• Maksym
• Eerik
• Christian
• Carlos
• Raul