Matthias Wolff

Food intake monitoring: an acoustical approach to automated food intake activity detection and classification of consumed food

Obesity and nutrition-related diseases are currently growing challenges for medicine. A precise and timesaving method for food intake monitoring is needed. For this purpose, an approach based on the classification of sounds produced during food intake is presented. Sounds are recorded non-invasively by miniature microphones in the outer ear canal. A database of 51 participants eating seven types of food and consuming one drink has been developed for algorithm development and model training. The database is labeled manually using a protocol with introductions for annotation. The annotation procedure is evaluated using Cohens kappa coefficient. The food intake activity is detected by the comparison of the signal energy of in-ear sounds to environmental sounds recorded by a reference microphone. Hidden Markov models are used for the recognition of single chew or swallowing events. Intake cycles are modeled as event sequences in finite-state grammars. Classification of consumed food is realized by a finite-state grammar decoder based on the Viterbi algorithm. We achieved a detection accuracy of 83% and a food classification accuracy of 79% on a test set of 10% of all records. Our approach faces the need of monitoring the time and occurrence of eating. With differentiation of consumed food, a first step toward the goal of meal weight estimation is taken.

Voice characteristics conversion for TTS using reverse VTLN

In the past, several approaches have been proposed for voice conversion in TTS systems. Mostly, conversion is done by modification of the spectral properties and pitch to match a certain target voice. This conversion causes distortions that deteriorate the quality of the synthesized speech. In this paper we investigate a very simple and straightforward method for voice conversion. It generates a new voice from the source speaker instead of generating a certain target speakers voice. For application in TTS systems it is often sufficient to synthesize new voices that sound sufficiently different to be distinguishable from each other. This is done by applying a spectral warping technique that is commonly used for speaker normalization in speech recognition systems called vocal tract length normalization (VTLN). Due to the low requirements of resources this method is especially suited for embedded systems.

Statistical Classifiers for Structural Health Monitoring

Multisensor problems are important tasks in the field of structural health monitoring. By means of signals originating from different sensors, we have to make a decision about the test object. We describe a universal, largely problem independent method, which applies statistical classifiers in order to identify objects or assess their state. This work presents the results of a series of studies, which systematically investigated such approaches for a great variety of technical and biological signals. We give an overview on the theoretical background and describe two selected application examples.

Toward spontaneous speech Synthesis-utilizing language model information in TTS

State-of-the-art speech synthesis systems achieve a high overall quality. However, synthesized speech still lacks naturalness. To produce more natural and colloquial synthetic speech, our research focuses on integration of effects present in spontaneous speech. Conventional speech synthesis systems do not consider the probability of a word in its context. Recent investigations on corpora of natural speech showed that words that are very likely to occur in a given context are pronounced less accurately and faster than improbable ones. In this paper three approaches are introduced to model this effect found in spontaneous speech. The first algorithm changes the speaking rate directly by shortening or lengthening the syllables of a word depending on the language model probability of that word. Since probable words are not only pronounced faster but also less accurately this approach was extended by selecting appropriate pronunciation variants of a word according to the language model probability. This second algorithm changes the local speaking rate indirectly by controlling the grapheme-phoneme conversion. In a third stage, a pronunciation sequence model was used to select the appropriate variants according to their sequence probability. In listening experiments test participants were asked to rate the synthesized speech in the categories colloquial impression and naturalness. Our approaches achieved a significant improvement in the category colloquial impression. However, no significantly higher naturalness could be observed. The observed effects will be discussed in detail.

Speech synthesis using stochastic Markov graphs

Speech synthesis systems basing on concatenation of natural speech segments achieve a high quality in terms of naturalness and intelligibility. However, in many applications such systems are not easy to apply because of the huge demand for storage capacity. Speech synthesis systems based on HMMs could be an alternative to concatenative speech synthesis systems but do not yet achieve the quality needed for use in applications. In one of our research projects we investigate the possibility of combining speech synthesis and speech recognition to a unified system using the same databases and similar algorithms for synthesis and recognition. In this context we examine the suitability of stochastic Markov graphs instead of HMMs to improve the performance of such synthesis systems. The paper describes the training procedure we used to train the SMGs, explains the synthesis process and introduces an algorithm for state selection and state duration modeling. We focus particularly on issues which arise using SMGs instead of HMMs.

Analysis of verbal and nonverbal acoustic signals with the dresden UASR system

During the last few years, a framework for the development of algorithms for speech analysis and synthesis was implemented. The algorithms are connected to common databases on the different levels of a hierarchical structure. This framework which is called UASR (Unified Approach for Speech Synthesis and Recognition) and some related experiments and applications are described. Special focus is directed to the suitability of the system for processing nonverbal signals. This part is related to the analysis methods which are addressed in the COST 2102 initiative now. A potential application field in interaction research is discussed.

Classification of non-speech acoustic signals using structure models

Non-speech acoustic signals are widely used as the input of systems for non-destructive testing. In this rapidly growing field, the signals have an increasing complexity leading to the fact that powerful models are required. Methods like DTW and HMM, which are established in speech recognition, have been successfully used but are not sufficient in all cases. We propose the application of generalized structured Markov graphs (SMG). We describe a task independent structure learning technique which automatically adapts the models to the structure of the test signals. We demonstrate that our solution outperforms hand-tuned HMM structures in terms of class discrimination by two case studies using data from real applications.

Food intake recognition conception for wearable devices

Obesity is a growing healthcare challenge in present days. Objective automated methods of food intake monitoring are necessary to face this challenge in future. A method for non-invasive monitoring of human food intake behavior by the evaluation of chewing and swallowing sounds has been developed. A wearable food intake sensor has been created by integrating in-ear microphone and a reference microphone in a hearing aid case. A concept for food intake monitoring requiring low computational cost is presented. After the detection of food intake activity periods, signal recognition algorithms based on Hidden Markov Models distinguish several types of food based on the sound properties of their chewing sounds. Algorithms are developed using manual labeled records of the food intake sounds of 40 participants.

Auscultatory Blood Pressure Measurement using HMMs

This paper reports on a study of applying an HMM-based labeler along with a tailored feature extraction to Korotkoff sounds. These sounds can be heard through a stethoscope during the auscultatory blood pressure measurement usually done at medical practices. While this method works well when the patient is at rest, interfering noise from muscles and joints cause major problems when the subject is doing any activities like sports or fitness exercises. We propose a signal processing and classification method to overcome these difficulties and present first promising results.

Automatic Decision Making in SHM Using Hidden Markov Models

Decision making and classification methods are very important in the field of structural health monitoring and life cycle prediction. We want to introduce an approach basing on sequence classifiers which can be used to several applications without any explicit knowledge of structures. To illustrate the concept we explain the method by means of a special example. So we can demonstrate our approach detailed, but although not too abstract.