Kristian Kroschel

Karlsruhe Institute of Technology | KIT · Communications Engineering Lab (CEL)

Breathing or external respiration is based on the expansion and contraction of the lungs. This mechanical activity can be picked up at appropriate locations on the surface of the skin by a laser Doppler vibrometer. To extract the breathing signal from the raw vibrometer signal, a low-pass filter is required since the bandwidth of the breathing sign...

Four sources of data have been picked up following the conditions of the ethics commission: from students in a dormitory in the framework of a thesis, from 50 patients with diseases in the cardiological department in a hospital, some test samples from premature infants in a hospital for children and a few more at the IOSB, a Fraunhofer Institute. T...

In cases when the activity of the heart has to be observed without contact, the electrocardiogram or ECG is replaced by the vibrocardiogram or VCG. The VCG is gained by filtering the output signal of a laser Doppler vibrometer. Whereas the ECG describes the electrical operation of the heart, the VCG originates from its mechanical operation. Since i...

This book presents recent outcomes of the collaborative “Tricorder” project, which brings together partners from industry, research institutes and hospitals to deliver an easy contactless alternative for electrocardiograms (ECG). Featuring contributions investigating the possible applications of laser Doppler vibrometry (LDV) signals for the remote...

Anomaly detection plays in many fields of research, along with the strongly related task of outlier detection, a very important role. Especially within the context of the automated analysis of video material recorded by surveillance cameras, abnormal situations can be of very different nature. For this purpose this work investigates Generative-Adve...

Background: The vibrocardiography (VBCG) is a laser-based technique to monitor the heart rhythm without any contact to the body. The aim of this study was to evaluate whether the VBCG is able to detect the vibration patterns of the atria. Methods: Simultaneous recordings of the ECG and VBCG in two cohorts were evaluated. Results: The VBCG deli...

Zusammenfassung Laser-Doppler-Vibrometer (LDV) messen kleinste Schwingungsamplituden und haben neben technischen auch biomedizinische Anwendungen wie zum Beispiel die Schwingungsmessung auf der Haut. Die Überwachung des Herzrhythmus wird aktuell mit dem Elektrokardiogramm (EKG) durchgeführt, wobei Elektroden auf der Haut befestigt werden. Eine berü...

The performance of speech recognition systems decreases dramatically in noisy environments. A robust human-computer interaction system should therefore make use of both, acoustic and visual signals. In this paper we present an automatic vowel recognition system which can perform in quasi real-time. We focus mainly on the recognition of 5 different...

Laser Doppler Vibrometry (LDV) is known to be a possible diagnosis tool for many cardiac applications as the detection and monitoring of some important vital parameters (Heart Rate, Heart Rate Variability, Pulse Wave Velocity) in a non-contact and non-intrusive way. The technique has become known as Optical Vibrocardiography (VCG) i.e. by measuring...

We extend our work on an integrated object-based system for saliency-driven overt attention and knowledge-driven object analysis. We present how we can reduce the amount of necessary head movement during scene analysis while still focusing all salient proto-objects in an order that strongly favors proto-objects with a higher saliency. Furthermore,...

Multimodal attention is a key requirement for humanoid robots in order to navigate in complex environments and act as social, cognitive human partners. To this end, robots have to incorporate attention mechanisms that focus the processing on the potentially most relevant stimuli while controlling the sensor orientation to improve the perception of...

In this contribution four different far infrared sensor setups with different optical configurations are evaluated based on their performance for pedestrian detection. The focus of the measurements is on the impact of resolution and sensitivity on the detection performance. To overcome the expensive and time consuming process of ground truth genera...

In this paper an approach for dynamic camera selection in large video-based sensor networks for the purpose of multi-camera object tracking is presented. The sensor selection approach is based on computational geometry algorithms and is able to determine task-relevant cameras (camera cluster) by evaluation of geometrical attributes, given the last...

Zusammenfassung In diesem Artikel wird ein generischer Ansatz und eine Prozessarchitektur zur auftragsorientierten Sensordatenanalyse vorgestellt. Zunächst werden anhand von Anwendungsbeispielen aus der Videoüberwachung die Vorteile einer auftragsorientierten Informationsverarbeitung gegenüber dem klassischen sensororientierten Ansatz beschrieben....

In this paper a task-oriented approach for object trackingin large distributed camera networks is presented. Thiswork includes three main contributions. First a generic processframework is presented, which has been designed fortask-oriented video processing. Second, system componentsof the task-oriented framework needed for the task of multicamerap...

The opto-acoustic scene analysis is an extremely important as well as a challenging task for a humanoid robot. By the opto-acoustic scene analysis, the guided and autonomous exploration of the environment by means of acoustic and/or visual perception is meant. On the one hand, the perception ability is necessary to interact with humans in a humanoi...

Performance of perceptive systems depends on the quality of the input data. In this contribution, an approach to evaluate perception performance as a function of quality of the sensor data is presented. Standardized quality metrics support the imaging sensors performance measurement. Several imaging setups are analyzed with real world experiments....

Besides resolution, an important performance parameter of a FIR camera is the sensitivity. It depends on the sensitivity of the detector array itself and the characteristics of the optic. The effects of the optic are considerably driven by the f-number, with high values resulting in decreased sensitivity, but providing the possibility for simple le...

The performance of perceptive systems depends on a large number of factors. The practical problem during development is, that this dependency is very often not explicitly known. In this contribution we address this problem and present an approach to evaluate perception performance, as a function of e.g. quality of the sensor data. The approach is t...

In this paper an approach for dynamic sensor selection in large video-based sensor networks for the purpose of multi-camera object tracking is presented. The sensor selection approach is based on computational geometry algorithms and is able to determine task-relevant cameras (camera cluster) by evaluation of geometrical attributes, given the last...

In this paper, a robust and efficient approach for multicamera human tracking is presented. The approach is integrated in an experimental surveillance system, based on a camera network with a task-oriented architecture. At sensor level, image processing algorithms are applied for object detection and feature extraction. Additionally, for each objec...

In this paper, an agent-based software architecture for automated wide area video surveillance systems is presented. The proposed concept is designed for detection and tracking of moving objects across multiple camera views. The surveillance system consists of a decentralized collaborative sensor network with object- and task-oriented architecture....

The lack of publicly available annotated databases is one of the major barriers to research advances on emotional information processing. In this contribution we present a recently collected database of spontaneous emotional speech in German which is being made available to the research community. The database consists of 12 hours of audio-visual r...

In real world scenes a large variety of objects can occur, that have to be recognized and localized by humanoids. In this paper we propose a new approach to initialize three dimensional pose and shape parameters for a large variety of objects, which is applicable to single images. In order to feed the recognition system with a priori knowledge, thr...

Emotion primitive descriptions are an important alternative to classical emotion categories for describing a human’s affective expressions. We build a multi-dimensional emotion space composed of the emotion primitives of valence, activation, and dominance. In this study, an image-based, text-free evaluation system is presented that provides intuiti...

We present novel methods for estimating spontaneously expressed emotions in speech. Three continuous-valued emotion primitives are used to describe emotions, namely valence, activation, and dominance. For the estimation of these primitives, support vector machines (SVMs) are used in their application for regression (support vector regression, SVR)....

This paper deals with the problem of decision making in the context of forward collision mitigation system design. The authors present a multilevel collision mitigation (CM) approach that allows a flexible tradeoff between potential benefit and the risk associated with driver acceptability and product liability. Due to its practical relevance, algo...

This paper presents an efficient approach for the combined audio-visual localization of a fallen object. The localization algorithm is incorporated into a multi-sensor robotic platform to supervise and to adapt dynamically the discrete plan of a robot performing a pick-and-place task. Initial audio position estimates of the fallen object are delive...

Collision mitigation systems are important driver assistance systems which promise to improve traffic safety in terms of reducing the number of road accidents and accident severity. Increasing research effort has been devoted to these systems within the past few years and first systems have recently been introduced into the market. However, these s...

This paper deals with the problem of decision making in the context of forward collision mitigation system design. We present a multi-level collision mitigation approach which allows a flexible tradeoff between potential benefit and the risk associated with driver acceptability and product liability. Due to its practical relevance, we further outli...

This paper brings together two current trends in emo-tion recognition: feature-based categorical classification and primitives-based dynamic emotion estimation. In this study, listeners rated a database of acted emotions using the three-dimensional emotion primitive space of valence, activation, and dominance. The emotion primitives were estimated...

In this paper, we consider both speaker dependent and listener dependent aspects in the assessment of emotions in speech. We model the speaker dependencies in emotional speech produc-tion by two parameters which describe the individual's emo-tional expression behavior. Similarly, we model the listener's emotion perception behavior by a simple param...

Angry -0.49 0.51 0.65 -0.34 0.39 0.49 -0.23 0.47 0.47 Happy 0.58 0.44 0.27 0.23 0.13 0.11 0.12 -0.08 -0.01 Neutral -0.18 -0.33 -0.08 -0.15 -0.26 -0.14 -0.16 -0.36 -0.20 Sad -0.49 -0.70 -0.61 -0.31 -0.47 -0.42 -0.48 -0.53 -0.58 Standard deviations Speaker 1 Speaker 2 Speaker 3 Val Act Dom Val Act Dom Val Act Dom Angry 0.23 0.27 0.21 0.13 0.12 0.11 0...

In this paper, we present the concept of a collision prevention assistant, a system that we believe can significantly contribute to road safety. We propose a new situation assessment algorithm which is tailored to the action of braking and that further accounts for the nonlinearities that arise when vehicles cut out or come to a standstill. The eff...

Time delay estimates in microphone pairs are commonly used for acoustic sound source localization. In real environments, the reliability of time delay estimates quickly degrades for increasing noise or room reverberation levels. The paper presents two computationally non-demanding criteria to determine the reliability of time delay estimates. The c...

Decentralized Kalman lters are often used in multi-sensor target tracking as such a distributed fusion architecture has several ad- vantages compared with centralized ones. On the other hand, dis- tributed fusion is not only conceptually more complex but the re- quired bandwidth is also likely to be a lot higher. However, a trade- off between bandw...

Emotion recognition in speech is an important research objective in the field of man-machine interfaces. This paper focuses on human labeling of emotions to create training data for emotion recognition systems. A three-dimensional emotion space concept is used to address the complexity of emotions in natural speech. As an evaluation tool, an iconic...

A humanoid robot must be able to cooperate efficiently and safely with humans in an unconstrained environment. Furthermore, it should not only present no danger to humans itself, but also be able to detect dangerous situations for the user. To restrict the processing only to salient objects and events, attention systems modeling the human visual sy...

The classical linearized conversion of measurements from polar or spherical coordinates to Cartesian ones generates a bias restricting the use of this conversion to cases where the bias can be neglected. In this work, the validity limits for the classical 2D transformation from polar to Cartesian coordinates, as derived in previous work, are shown...

Decentralized Kalman Filters are often used in multi-sensor target tracking as such a distributed fusion architecture has several advantages compared with centralized ones. On the other hand, distributed fusion is not only conceptually more complex but the required bandwidth is also likely to be a lot higher. This is especially true if the final es...

Time Difference Of Arrival (TDOA) estimates are used for pas-sive acoustic single sound source localization with microphone arrays. The technique of choice in most systems for TDOA esti-mation is the Generalized Cross-Correlation (GCC) method [1]. For a multi-sound source scenario, cross-correlation terms of the active sound source signals as well...

Microphone arrays can be employed for passive acoustic source lo-calization using time difference of arrival (TDOA) estimates in microphone pairs. The most common method for this TDOA estimation is the generalized cross-correlation (GCC) method which is also used in this work. In an office room en-vironmental noise influences and reverberation effe...

This paper presents some new results that prove the usability of a Surface Mountable Sub-Wavelength (SMSW) Array for low-cost, hands-free applications and as a front-end for speech recognition systems. A noise suppression system using an SMSW-array shows a strong directivity index and improves the recognition rates of the speech recognizer used for...

In this publication a method for tracking a speaker with acoustical information by means of a microphone array is presented. A sound source localization algorithm based on the time delays of arrival of sound waves in microphone pairs provides initial position estimates. These significantly varying estimates are spatially filtered by an adaptive Kal...

Decentralized Kalman Filters are often used in multi-sensor target tracking as such a distributed fusion architecture has several advantages compared with centralized ones. On the other hand, distributed fusion is not only conceptually more complex but the required bandwidth is also likely to be a lot higher, especially if the nal estimate shall be...

Die bei den bisherigen Betrachtungen gemachten Voraussetzungen zur Lösung des einfachen und multiplen Parameterschätzproblems sollen nun beim Entwurf des Schätzers um eine Annahme erweitert werden: In diesem Kapitel sollen lineare Schätzeinrichtungen untersucht werden. Linear bedeutet, dass der Schätzvektor \(\hat q\) eine Linearkombination des ges...

Die statistische Informationstechnik befasst sich mit so unterschiedlichen Aufgaben wie der Extraktion von verrauschten Daten, die über einen gestörten Kanal übertragen wurden, der Erkennung gesprochener Wörter, der Schätzung der Impulsantwort eines unbekannten Übertragungskanals unter Störeinfluss, der Vorhersage des Signalverlaufs zum Laufzeitaus...

In diesem Kapitel soll auf die Realisierung von Systemen eingegangen werden, die zur Signalerkennung und Mustererkennung dienen. Synonyme Bezeichnungen für diese Systeme sind im Falle der Signalerkennung Empfänger, insbesondere Optimalempfänger, bei der Mustererkennung wird der Begriff Klassifikator verwendet. Üblich sind auch die Bezeichnungen Sch...

Nachrichtenquellen für Datensignale, Sprache, Bilder usw. liefern zwei oder mehrere Ereignisse M i , i = 1,..., M, denen nach den Ergebnissen des vorangehenden Kapitels die Vektoren si sowie die Auftritts- oder A-priori-Wahrscheinlichkeiten P i zur Beschreibung des jeweiligen Sendesignals zugeordnet werden können. Am Ausgang des Vektorkanals steht...

In diesem Kapitel sollen die wesentlichen Begriffe der Statistik und der Systemtheorie aufgeführt werden, soweit man sie im Rahmen dieses Buches benötigt. Dabei wird auf Vollständigkeit und ausführliche Behandlung verzichtet und auf die Literatur, z.B. [Pap65], [Kre68], [Han83], verwiesen.

A system for three-dimensional passive acoustic speaker localization and tracking using a microphone array is presented and evaluated. Initial speaker position estimates are provided by a time-delay-based localization algorithm. These raw estimates are spatially smoothed by a multiple model adaptive estimator consisting of three extended Kalman fil...

In this publication a method for tracking a speaker with acoustical information by means of a microphone array is presented. A sound source localization algorithm based on the time delays of arrival of sound waves in microphone pairs provides initial position estimates. These significantly varying estimates are spatially filtered by an adaptive Kal...

Time Difference Of Arrival (TDOA) estimates can be used for passive acoustic multiple sound source localization with microphone arrays. The TDOA estimation is based on the cross-correlation function of two signals of a microphone pair. Existing systems assume only one dominant sound source per analysis frame rejecting the localization infor-mation...

In Abschnitt 3.5 wurde gezeigt, daß man die Übertragungseigenschaften linearer digitaler Systeme im z-Bereich durch eine gebrochen rationale Übertragungsfunktion beschreiben kann. Durch die Wahl der Parameter b µ , und a v wird es möglich, bestimmte Eigenschaften bezüglich des Amplitudenganges und der Phase bzw. der Gruppenlaufzeit zu realisieren....

Die moderne Meßtechnik macht seit einer Reihe von Jahren intensiven Gebrauch von den Methoden der digitalen Signalverarbeitung. Mit der Entwicklung sehr schneller digitaler Bausteine gilt dies in immer stärkerem Maße auch für die Echtzeitsignalanalyse mit den Mitteln der schnellen Fourier-Transformation. Traditionelle analoge Spektralanalysatoren a...

In der Systemtheorie zeitkontinuierlicher Signale verwendet man zur Transformation des Signals x(t) in den Frequenzbereich die Laplace-Transformation anstelle der Fourier-Transformation immer dann, wenn Konvergenzprobleme des Fourier-Integrals auftreten. Schon bei einer elementaren Funktion wie der kontinuierlichen Sprungfunktion treten diese Konve...

Diskrete Signale erhält man meist dadurch, daß man kontinuierliche Signale abtastet. Beispiele für kontinuierliche Signale sind die Spannung x к (t) mit dem zeitkontinuierlichen Parameter t am Ausgang eines Sensors oder die Grauwerte x к(u, υ) einer zweidimensionalen Bildvorlage mit den ortskontinuierlichen Parametern u und υ. Im Rahmen dieses Buch...

Es ist ein Funktionsgenerator zu programmieren, der ein Nutzsignal der Frequenz f 0 erzeugt und bei einer festen Abtastfrequenz f A arbeitet. Dabei sollen ein Sinus- und ein Sägezahnsignal generiert werden.

Das Spektrum eines stochastischen Prozesses ist durch die Wiener-Khintchine Beziehung (2.7.1) erklärt. Sie verknüpft die Autokorrelationsfolge und die spektrale Leistungsdichte über die Fourier-Transformation. Die Spektralschätzung wird somit eng im Zusammenhang mit der Schätzung der Autokorrelationsfolge stehen.

Alle traditionellen Verfahren der Spektralschätzung basieren auf der Annahme, daß die in der Messung erfaßte Musterfunktion außerhalb des Beobachtungsintervalls verschwindet. Implizit wird also stets eine Fensterung im Zeitbereich vorgenommen, was gleichbedeutend mit dem Ansatz einer Autokorrelationsfolge endlicher Länge ist. Der geschätzte Ausschn...

Die im vorangegangenen Kapitel behandelten rekursiven Filter sind in ihrem Systemverhalten und bezüglich der Entwurfsmöglichkeiten eng mit den klassischen analogen Filtern verwandt. Beide Systemtypen werden durch Pole und Nullstellen charakterisiert, die Impulsantworten sind theoretisch zeitlich unbegrenzt.

Der vorliegende Ansatz zeigt neue Methoden zur Vorhersage der Verfahrensleistung bei der Detektion von Mikrokalzifikationen in Mammogrammen auf. Dazu werden für ein gegebenes Mammogramm-Auswertungsverfahren einerseits für die zu detektierenden Mikrokalzifikationen, andererseits für den verbleibenden Hintergrund-Bildbereich separate Modelle erstellt...