Remote monitoring, distress detection by slightest invasive systems: Sound recognition based on hierarchical i-vectors

Abstract

Europe has a growing aging population, leading to the need for adapted healthcare services. Our work aims at proposing a solution for falls detection of elderly people using sound recognition based on a hierarchical i-vectors system. The system presented in this paper improves significantly the accuracy of sound recognition compared to the state of the art methods. The latter provides a good recognition rate of 81.98% on noiseless sounds. This system needs to be tested in a noisy environment and this can be improved by using new sound descriptors.

Smart Sound Sensor to Detect the Number of People in a Room

“…We have already developed a smart audio sensor [1] allowing the detection of distress situations but also of people activity through sound environment analysis. This proposed sensor analyse the sound environment and is able using different techniques to recognize 18 sound classes in a continuous audio flow.…”