George Papanikolaou

Novel wavelet domain Wiener filtering de-noising techniques: Application to bowel sounds captured by means of abdominal surface vibrations

Abstract This work focuses on the design and evaluation of efficient and accurate de-noising algorithms that combine robust signal enhancement and minimum signal distortion. The proposed method introduces novel, frequency depended, parametric, Wiener filtering techniques that involve Discrete Wavelet Transform and Wavelet Packets. Implementations of various decomposition schemes, different mother wavelets and various thresholding options were tested, while perceptual criteria were also taken into account. The introduced de-noising approach has been extensively tested on human bowel sounds, captured by means of abdominal surface vibration recordings, in order to be further utilized as a diagnostic tool. Qualitative and quantitative analysis of the methods performance, when applied to various types of recorded and synthetic sounds, revealed that the new approach works excellent with favourable results.

Joint wavelet video denoising and motion activity detection in multimodal human activity analysis: application to video-assisted bioacoustic/psychophysiological monitoring

The current work focuses on the design and implementation of an indoor surveillance application for long-term automated analysis of human activity, in a video-assisted biomedical monitoring system. Video processing is necessary to overcome noise-related problems, caused by suboptimal video capturing conditions, due to poor lighting or even complete darkness during overnight recordings. Modified wavelet-domain spatiotemporal Wiener filtering and motion-detection algorithms are employed to facilitate video enhancement, motion-activity-based indexing and summarization. Structural aspects for validation of the motion detection results are also used. The proposed system has been already deployed in monitoring of long-term abdominal sounds, for surveillance automation, motion-artefacts detection and connection with other psychophysiological parameters. However, it can be used to any video-assisted biomedical monitoring or other surveillance application with similar demands.

Live Broadcasting of High Definition Audiovisual Content Using HDTV over Broadband IP Networks

The current paper focuses on validating an implementation of a state-of-the art audiovisual (AV) technologies setup for live broadcasting of cultural shows, via broadband Internet. The main objective of the work was to study, configure, and setup dedicated audio-video equipment for the processes of capturing, processing, and transmission of extended resolution and high fidelity AV content in order to increase realism and achieve maximum audience sensation. Internet2 and GEANT broadband telecommunication networks were selected as the most applicable technology to deliver such traffic workloads. Validation procedures were conducted in combination with metric-based quality of service (QoS) and quality of experience (QoE) evaluation experiments for the quantification and the perceptual interpretation of the quality achieved during content reproduction. The implemented system was successfully applied in real-world applications, such as the transmission of cultural events from Thessaloniki Concert Hall throughout Greece as well as the reproduction of Philadelphia Orchestra performances (USA) via Internet2 and GEANT backbones.

Computer‐aided systems for prolonged recording and analysis of human bowel sounds

This paper presents the design of noninvasive systems for the investigation of the small bowel motility patterns. The basic idea was to implement dedicated devices for the recording and analysis of human bowel sounds. Properly modified transducers were used for the caption of gastrointestinal sound signals. A data acquisition ambulatory system, consisting of a preamplifier and a recording unit, was initially designed capable for maximum continuous recording of 6 h. After the first experimental results with this system, the necessity for longer recording duration was obvious. For this reason the installation of a computer‐based stationary system in a special reformed examination room was the next development step. Triggering of the input data and rejection of the regions with absence of gastrointestinal activity allowed much more recording duration. A computerized signal processing system, which involves equalizing procedures, noise cancellation techniques, and neural network algorithms, was used for the o...

Machine Learning Algorithms for Environmental Sound Recognition: Towards Soundscape Semantics

This paper investigates methods aiming at the automatic recognition and classification of discrete environmental sounds, for the purpose of subsequently applying these methods to the recognition of soundscapes. Research in audio recognition has traditionally focused on the domains of speech and music. Comparatively little research has been done towards recognizing non-speech environmental sounds. For this reason, in this paper, we apply existing techniques that have been proved efficient in the other two domains. These techniques are comprehensively compared to determine the most appropriate one for addressing the problem of environmental sound recognition.

Embedding sound localization and spatial audio interaction through coincident microphones arrays

This paper discusses a methodology for embedding sound localization techniques for spatial audio interaction, aiming at matching the low computing capabilities of mobile and embedded systems. The main goal is to implement a sound localization system, using a microphone array that combines increased accuracy with compromised computational load and applicable layout size. In particular, four cardioid microphones are placed in a cross-shape arrangement, thus forming a planar coincident microphones array for horizontal direction of arrival estimation. The incorporation of two additional microphones at the perpendicular plane is also considered for 3D audio localization. The implemented system is evaluated through simulation experiments and real-world field measurements in comparison to B-Format based localization. Joint time frequency analysis is considered for improving the localization accuracy in pure SNR conditions. The utilization of multiple arrays is also discussed for 2D and 3D position estimations, as well as signal enhancement by means of time-delay compensation.

Mobile Audio Intelligence: From Real Time Segmentation to Crowd Sourced Semantics

The task of general audio detection and segmentation based in means of machine learning is very popular and high-demanding procedure nowadays. Most relevant works in the last decade aim at modelling audio in order to conduct a semantics analysis and a high--level categorization. A generic strategy that would detect audio events as means of transitions from one audio state to another is considered interesting and would support whole classification workflow. This work investigates the possibilities in designing a robust bimodal segmentation algorithm for audio that would perform well in different conditions without relying on complicated machine learning schemes by minimizing prior knowledge for detection model, and thus, delivering consistent performance for any input signal and computing environment. Additionally, a modern user-generated content approach for populating and updating ground truth databases is presented. Both techniques are implemented and embedded as upgrades, in a mobile software environment for smartphones.

Adaptive phoneme alignment based on rough set theory

The current work describes a phoneme matching algorithm based on rough set concepts. The objective of this type of algorithms is focused on the localization of the phonemic content of a specific spoken occurrence. According to the proposed algorithm, a number of rough sets containing the multiple expected phonemic instances in a sequence are created, each defined by a set of short term frames of the voice signal. The properties of the corresponding information system are derived from a features set calculated from the speech signal upon initiation. Given the above, an iterative procedure is applied by updating the phoneme instances versus the optimization of the accuracy metric. The main advantage of this algorithm is the absence of a training phase allowing for wider speaker adaptability and independency. The current paper focuses on the feasibility of the task as this work is still in early research stage.

New design concepts for the construction of an omnidirectional sound source

This paper presents a study for the optimum design and construction of an omnidirectional point source. The point source model of theoretical acoustics is used as an elementary unit for electroacoustic measurements in the area of room and free‐field acoustics. The spherical radiation requirement for the entire acoustic spectrum and the finite dimensions of the point source model contradict the construction in practice. This is because of high‐frequency beaming and the existing dimensions of radiators. A new design approach of using horn loaded cone loudspeakers and properly placed high‐frequency diffusers was used to minimize these problems. Twelve radiators loaded on pentagon‐shaped horns were placed in a dodecahedron topology to build a practical point source with a useful frequency response of 80–16 000 Hz and an omnidirectional radiation pattern up to 12 000 Hz. The results display improvement in the radiation pattern compared to those of the already existing point sources, due to the use of horns and diffusers. Increase in the sound pressure level is also displayed because of the obvious increase of the directivity index of each individual radiating element.

High accuracy block-matching sub-pixel motion estimation through detection of error surface minima

The present paper focuses on high-accuracy block-based sub-pixel motion estimation utilizing a straightforward error minimization approach. In particular, the mathematics of bilinear interpolation are utilized for the selection of the candidate motion vectors that minimize the error criterion, by estimating local minima in the error surface with arbitrary accuracy. The implemented approach favors optimum accuracy over computational load demands, making it ideal as a benchmark for faster methods to compare against; however, it is not best suited to real-time critical applications (i.e. video compression). Other video processing needs relying on motion vectors and requiring high-resolution/accuracy can also take advantage of the proposed solution (and its simplified nature in terms of underlying theoretical complexity), such as motion-compensation filtering for super resolution image enhancement, motion analysis in sensitive areas (e.g. high-speed video monitoring, medical imaging, motion analysis in sport science, big-data visual surveillance, etc.). The proposed method is thoroughly evaluated using both real video and synthetic motion sequences from still images, adopting well-tested block-based motion estimation evaluation procedures. Assessment includes comparisons to a number of existing block-based methods with respect to PSNR and SSIM metrics over ground-truth samples. The conducted evaluation takes into consideration both the original (arbitrary-accuracy) and the truncated motion vectors (after rounding them to the nearest half, quarter, or eighth of a pixel), where superior performance with more accurate motion vector estimation is revealed. In this context, the degree to which sub-pixel motion estimation methods actually produce sub-pixel motion vectors is investigated, and the implications thereof are discussed.