Bogdan Orza

Fuzzy intensification operator based contrast enhancement in the compressed domain

With the increasing sizes of high resolution images, their storage and processing directly in the compressed domain has significantly gained importance. Algorithms for compressed domain image processing provide a powerful computational alternative to classical (pixel level) based implementations. While linear algorithms can be applied straightforward to the JPEG compressed images, this is not the case for nonlinear image processing, as for example contrast enhancement algorithms. In this paper a new implementation in the compressed domain of a very efficient contrast enhancement, based on fuzzy set modeling and on a fuzzy intensification operator, is presented. The fuzzy set parameters are adaptively chosen by analyzing the statistics of the image data in the compressed domain, in order to optimally enhance the image contrast. The nonlinear enhancement procedure requires a grey level threshold, for which an adaptive implementation, taking into account the frequency content of each coefficient block in the DCT (Discrete Cosine Transform) encoded JPEG image is proposed. This guarantees the optimal quality at minimum computational cost. The experimental results for a set of various contrast images validate the good performance and functionality of the proposed implementation.

Advanced human-computer interaction in external resource annotation

This paper presents innovative multimedia annotation methods based on advanced human-computer interaction and (a)synchronous interpersonal collaboration features. They have been implemented within ERA (External Resource Annotation) framework with the scope of going beyond the existing solutions by introducing new annotation types using external resources. Thus, anyone can easily annotate the own resources with additional information such as real-time handwriting, text and animation, or even audio/video sequences stored into public data repositories, or recorded ad-hoc, in order to create value-added interactive content. ERA framework also provides an advanced rendering tool for such complex multimedia resources.

Blended Learning Concept and its Applications to Engineering Education

The technical educational act means theory and practice, individual study, group-based projects or experimental work that involves equipment, simulation/emulation software packages and laboratory applications. In order to develop advanced e-learning tools and integrate them within online platforms for higher and postgraduate engineering education, new methodologies should be taken into consideration: project-and problem based learning, virtual laboratory (remote access to laboratory infrastructure and task evaluation) or remote assistance for diploma projects and mobility grants. This paper presents new blended learning concepts and approaches and the manner they can be adopted in engineering education, by using the symbiotic computing paradigms adapted to blended learning, especially blended learning gap and blended learning bridge.

Computer-Supported Collaborative Work and Its Application to E-Health

Most of the articles published in the last years present the asynchronous collaborative components integrated into the E-Health complex applications. They do not refer to the synchronous collaboration among the medical staff. A possible approach is because E-Health systems are patient-centered and the physicians, nurses and paramedics have to collaborate in a transparent manner not being focused on the synchronous collaborative tools. Our proposal is focused on advanced concepts for improving the telemedicine services and the real time collaboration among the medical staff. The article is based on a virtual collaboration prototype that supports both interpersonal and inter-process collaborative services and can be considered the core of a virtual healthcare community.

Blind speech segmentation using spectrogram image-based features and Mel cepstral coefficients

This paper introduces a novel method for blind speech segmentation at a phone level based on image processing. We consider the spectrogram of the waveform of an utterance as an image and hypothesize that its striping defects, i.e. discontinuities, appear due to phone boundaries. Using a simple image destriping algorithm these discontinuities are found. To discover phone transitions which are not as salient in the image, we compute spectral changes derived from the time evolution of Mel cepstral parametrisation of speech. These socalled image-based and acoustic features are then combined to form a mixed probability function, whose values indicate the likelihood of a phone boundary being located at the corresponding time frame. The method is completely unsupervised and achieves an accuracy of 75.59% at a −3.26% over-segmentation rate, yielding an F-measure of 0.76 and an 0.80 R-value on the TIMIT dataset.

A compact robotic head for automatic acquisition of indoor panoramic images

Indoor exploration based on panoramic images has been of interest in both navigation and visualization scenarios of mobile intelligent systems, becoming an appealing research topic in hardware design and computational algorithms. A detailed profile of an indoor scene further used for 3D reconstruction purposes or visualization requires an accurate high-resolution rendering from multiple camera shots. In panoramic imaging, 360 degree Gigapixel panoramas involve a process of capturing multiple segments at a predefined step according to the physical and optical characteristics of the camera (focal length, field of view, angle of view). Manual shooting in such scenes becomes cumbersome, as high precision and stability is required for prolonged time. The current work, proposes a hardware implementation and software control of a panoramic robotized head (GoSphere), used to automatically capture panoramic segments, that further are to be stitched into panoramic renderings. The proposed model features a light-weight portable design for action cameras, allowing a two axes control for a 360 × 180 degree coverage of the scene. Experimental validation has been made in comparison with a standard panoramic capturing assembly. Limitations and improvement proposals are presented for further development and extensions.

Blended Assessment Concepts for Formal and Informal Engineering Education

Multiple assessment options have been adopted in engineering education in the last years due to different reasons. Most of them are related to increasing the efficiency of the evaluation process applied in each activity performed by students/trainees and decreasing stress among them. The assessment methodology implemented in the formal and informal education programs mentioned in the article includes the automated evaluation approach consisting of assessment of conceptual/theoretical knowledge, as well as competences achieved within individual and/or team projects, hands-on laboratory, or simulation work. An analysis of the assessment scores reveals that end-of-programs assessments looks better if continuous evaluation methods are applied during the whole study act while students and trainees develop their own modes of learning and working based on the assessment results.

Use of Machine Learning for Improvement of Similarity Searches of Patients

Due to various factors (human error, broken communication links) patient information within a healthcare enterprise varies for a given patient. Medical personnel and technicians are often faced with the problem of searching similar patients within the hospital information system. The heterogeneity of attributes on which the search is performed increases the difficulty of retrieving conclusive results, thus simple search rules are unsatisfactory in this case.

The SWARA speech corpus: A large parallel Romanian read speech dataset

This paper introduces one of the largest Romanian speech datasets freely available for both academic and commercial use. The dataset comprises speech data recorded over the last year from 12 speakers, along with 5 other speakers previously recorded in a separate environment. The data was manually segmented at utterance-level and semi-automatically labelled at phone-level. The resulting corpus amounts to approximately 21 hours of high-quality read speech data, split into over 19,000 utterances. The speakers read between 921 and 1493 utterances each. 880 utterances are common to all speakers and add up to over 16 hours of parallel data. We present the steps of performing the recordings and data segmentation, as well as a first use of this corpus in the context of synthetic voice development.

A fuzzy logic approach for the fast approximate computation of image transforms from block JPEG DCT coefficients

Computationally efficient implementation of different image processing and analysis algorithms is crucial in many modern real world applications. Several modern visual analytics applications involve processing large amounts of compressed image/video data on restricted hardware resources, in a numerically efficient manner. In this context, the formulation of some image analysis/processing algorithms in the compressed image/video domain can bring significant benefits in the computational complexity reduction. An important step to image analysis, feature extraction, based on unitary image transforms is addressed in this paper, in a general form. Starting from the fact that in most popular video compression the intraframes are block discrete cosine transform (DCT) encoded, we present and propose a solution to computing any unitary image transform at pixels block level, starting from the DCT coefficients of the block. Since approximate forms of the transforms are often satisfactory, we propose a fuzzy logic system based approach to optimize the computational complexity with minimal loss of image information. The method is exemplified for the case of the Walsh transform, which shows that even the fastest implementation provides (by reconstruction) images of good quality.