Autonomous Machine Learning Framework for Detecting People Aliveness

Determining the aliveness of people is essential for various people safety systems. While the accuracy of the aliveness determination is a key concern, there exist technical challenges in the aliveness determination with high accuracy. Our approach to the challenges is to incorporate a set of effective design tactics into a software framework. We employ machine learning models in the detection process and apply the continuous optimization of the aliveness model using autonomous computing principles. This paper presents a comprehensive framework which consists of design and implementation. Our extensive experiments show that the accuracy of aliveness determination with this framework outperforms at least 30% of conventional approaches.