Jean-François Serignat

Information extraction from sound for medical telemonitoring

Today, the growth of the aging population in Europe needs an increasing number of health care professionals and facilities for aged persons. Medical telemonitoring at home (and, more generally, telemedicine) improves the patients comfort and reduces hospitalization costs. Using sound surveillance as an alternative solution to video telemonitoring, this paper deals with the detection and classification of alarming sounds in a noisy environment. The proposed sound analysis system can detect distress or everyday sounds everywhere in the monitored apartment, and is connected to classical medical telemonitoring sensors through a data fusion process. The sound analysis system is divided in two stages: sound detection and classification. The first analysis stage (sound detection) must extract significant sounds from a continuous signal flow. A new detection algorithm based on discrete wavelet transform is proposed in this paper, which leads to accurate results when applied to nonstationary signals (such as impulsive sounds). The algorithm presented in this paper was evaluated in a noisy environment and is favorably compared to the state of the art algorithms in the field. The second stage of the system is sound classification, which uses a statistical approach to identify unknown sounds. A statistical study was done to find out the most discriminant acoustical parameters in the input of the classification module. New wavelet based parameters, better adapted to noise, are proposed in this paper. The telemonitoring system validation is presented through various real and simulated test sets. The global sound based system leads to a 3% missed alarm rate and could be fused with other medical sensors to improve performance

Audio packet loss over IP and speech recognition

This paper deals with the effects of packet loss on speech recognition over IP connections. The performance of our continuous French speech recognition system is here evaluated for different transmission scenarios. A packet loss simulation model is first proposed in order to simulate different channel degradation conditions. The packet loss problem is also investigated in real transmissions through IP. Because packet loss impact may be different according to the speech coder used to transmit data, different transmission conditions with different audio codecs are also investigated. Several reconstruction strategies to recover lost information are then proposed, and tested. Another solution for dialog applications is also suggested, where the relative weight of the language and acoustic model is changed according to the packet loss rate. The results show that the speech recognition performance can be augmented by the solutions here presented.

Complete Sound and Speech Recognition System for Health Smart Homes: Application to the Recognition of Activities of Daily Living

This chapter presents the AUDITHIS system which performs real-time sound analysis from eight microphone channels in Health Smart Home associated to the autonomous speech analyzer RAPHAEL. The evaluation of AUDITHIS and RAPHAEL in different settings showed that audio modality is very promising to acquire information that are not available through other classical sensors. Audio processing is also the most natural way for a human to interact with his environment. Thus, this approach particularly fits Health Smart Homes that include home automation (e.g., voice command) or other high level interactions (e.g., dialogue). The originality of the work is also to include sounds of daily living as indicators to distinguish distress from normal situations. First development gave acceptable results for the sound recognition (72% correct classification) and we are working on the reduction of missed-alarm rate to improve performance in the near future. Although the current system suffers a number of limitations and that we raised numerous challenges that need to be addressed, the pair AUDITHIS and RAPHAEL is, to the best of our knowledge, one of the first serious attempts to build a real-time system that consider sound and speech analysis for ambient assisted living. This work also includes several evalu- ations on data acquired from volunteers in a real health smart home condition. Further work will include refinement of the acoustic models to adapt the speech recognition to the aged population as well as connexion to home automation systems.

Speech recognition in a smart home: Some experiments for telemonitoring

Because of the aging of the population, low-cost solutions are required to help people with loss of autonomy staying at home rather than in public health centers. One solution is to assist human operators with smart information systems. In this case, position and physiologic sensors already give important information, but there are few studies about the utility of sound in patients habitation. However, sound classification and speech recognition may greatly increase the versatility of such a system: this will be provided by detecting short sentences or words that could characterize a distress situation for the patient. Moreover, analysis and classification of sounds emitted in patients habitation may be useful for patients activity monitoring. In this paper, we present a global speech and sound recognition system that can be set-up in a flat. Eight microphones were placed in the Health Smart Home of Grenoble (named HIS, a real living flat of 47m2) to automatically analyze and classify different sounds and speech utterances (e.g.: normal or distress French sentences). Sounds are clustered in eight classes but this aspect is not discussed in this paper. For speech signals, an input utterance is recognized and a subsequent process classifies it in normal or distress, by analysing the presence of distress keywords. An experimental protocol was defined and then this system has been evaluated in uncontrolled conditions in which heterogeneous speakers were asked to utter predetermined sentences in the HIS. The results of this experiment, where ten subjects were involved, are presented. The Global Error Rate was 15.6%. Moreover, noise suppression techniques were incorporated in the speech and sound recognition system in order to suppress the noise emitted by known sources like TV or radio. An experimental protocol was defined and tested by four speakers in real conditions inside a room. Finally, we discuss the results of this experiment as a function of the noise source: speech or music.