Igor Vujović

Wavelet Quasi-Superresolution in Marine Applications

Possible marine applications of superresolution are in both military and civilian purposes, such as computer vision, target recognition, detection, identification, surveillance, synthetic aperture radar, etc. In this paper, a new algorithm for superresolution is proposed. The algorithm is based on wavelet theory and morphological structure of the image. Quasi-superresolution means usage of a single image for analysis instead of several. Wavelet quasi-superresolution can be applied with less complexity than standard superresolution, because wavelet interpolation of interlaced data and statistics of different images are skipped. New criterion for comparison of original and processed image is proposed. The result is single scalar which is easy to compare

Wavelet downsampling in image compression

Wavelet downsampling is a method of data compression. However, it is very possible that part of information might be lost when such method is used. These loses should be compensated by interpolation in reconstruction algorithm. Simple method of interpolation is used in this paper. Reconstruction algorithm was tested. The goal is to find out how much of downsampling specific data can take without undermining the purpose. For example, how much can be lost in medical image without change in diagnosis. The study can be repeated in different applications. We used the simplest wavelet - Haar. The results show that X-ray pulmonary images can be downsampled up to 10th level without change in diagnosis. In computer memory space, that means that image from size of approximately 4, 5 MB can be reduced to the size of approximately 35 KB or 128 times.

A novel approach for monitoring writing interferences during navigated transcranial magnetic stimulation mappings of writing related cortical areas.

BACKGROUND It has recently been shown that navigated repetitive transcranial magnetic stimulation (nTMS) is useful in preoperative neurosurgical mapping of motor and language brain areas. In TMS mapping of motor cortices the evoked responses can be quantitatively monitored by electromyographic (EMG) recordings. No such setup exists for monitoring of writing during nTMS mappings of writing related cortical areas. NEW METHOD We present a novel approach for monitoring writing during nTMS mappings of motor writing related cortical areas. COMPARISON WITH EXISTING METHOD(S) To our best knowledge, this is the first demonstration of quantitative monitoring of motor evoked responses from hand by EMG, and of pen related activity during writing with our custom made pen, together with the application of chronometric TMS design and patterned protocol of rTMS. RESULTS The method was applied in four healthy subjects participating in writing during nTMS mapping of the premotor cortical area corresponding to BA 6 and close to the superior frontal sulcus. The results showed that stimulation impaired writing in all subjects. The corresponding spectra of measured signal related to writing movements was observed in the frequency band 0-20 Hz. Magnetic stimulation affected writing by suppressing normal writing frequency band. CONCLUSION The proposed setup for monitoring of writing provides additional quantitative data for monitoring and the analysis of rTMS induced writing response modifications. The setup can be useful for investigation of neurophysiologic mechanisms of writing, for therapeutic effects of nTMS, and in preoperative mapping of language cortical areas in patients undergoing brain surgery.

Experimental results on visibility threshold in human visual perception for application in image/video coding and quality assessment

In assessment of image quality, as well as in image coding, end users are human beings. They need some type of image for work or entertainment. So it is logical to pursuit the goal of achieving better image quality by investigating what is better to human eyes. This paper is a contribution in that direction. Just-noticeable distortion and some aspects of it are discussed and experimental results on the visibility threshold of low gray tones are presented.

Multiresolution Approach to Processing Images for Different Applications

This book presents theoretical and practical aspects of the interaction between low and high level image processing. Multiresolution analysis owes its popularity mostly to wavelets and is widely used in a variety of applications. Low level image processing is important for the performance of many high level applications. The book includes examples from different research fields, i.e. video surveillance ; biomedical applications (EMG and X-ray) ; improved communication, namely teleoperation, telemedicine, animation, augmented/virtual reality and robot vision ; monitoring of the condition of ship systems and image quality control. Nowadays computers are expected to perform complex tasks involving the processing of huge amounts of data. People are often unaware that such intensive operations are required in computer vision tasks. Different phenomena related to pixel/voxel or global size can cause failure of higher vision applications. Such phenomena include illumination variations, noise, camera jitter, shadows, visibility, weather conditions, etc. This paper analyzes the influence of lower vi- sion tasks (e.g. denoising or thresholding) on higher vision tasks (e.g. motion segmentation or product quality). The interaction between lower and higher vision is illustrated with examples of visual quality control and advanced visualization in marine communications used to decrease the stress felt by seafarers due to their separation from their families.

Pre-Processing for Image Sequence Visualization Robust to Illumination Variations

Several images (a sequence) may be used to obtain better image quality. This method is perfect for super-resolution algorithms, which improve sub-pixel clarity of the image and allow a more detailed view. It is possible that illumination variations, e.g. those caused by a light source, lessen the benefits of super-resolution algorithms. The reduction of the quantity of such occurrences by stabilizing variations is important. An enhanced stabilization algorithm is proposed for purposes of reduction of variations in illumination. It is based on the energy contained in wavelet coefficients. In the proposed algorithm, energy plays a role of the memory buffer in memory-based techniques of illumination variation reduction. The benefits of the proposed image stabilization are the higher quality of images and better visualization. Possible applications are in surveillance, product quality control, engine monitoring, corrosion monitoring, micro/nano microscopy, etc.

Stabilising illumination variations in motion detection for surveillance applications

Since illumination variations may cause the misinterpretation of data for various higher vision applications and algorithms, this study aims to reduce such influence. In order to obtain a motion mask, which is input for a higher vision application, wavelet coefficients are calculated by applying two-dimensional lifting wavelet transform with two mother wavelets. Energy is calculated from the obtained wavelet coefficients. Morphological operations are used to improve output image. The developed algorithm is a robust algorithm further reducing false alarm readings caused by illumination variations (better false detection rate and percentage of correct classifications).

Corrosion Monitoring in Marine Environment Using Wavelet Description

The paper proposes a method, the so-called wavelet description, for corrosion monitoring in marine environment which is based on taking the energy difference contained in energy from coefficients derived by applying complex wavelet transforms. Then the rms value of coefficients has been accounted that form an illumination-independent energy representation of corroded and corrosion-free parts. The robustness on illumination variations has been achieved because by applying the proposed algorithm the energy representation enables reduction of the noise, since getting the difference of two subsequent images leaves only greater values of differences.

Nanotubes in NEMS -- Impact to the Marine Technology

Advance of technology touches most surely the area of marine engineering. Nanotechnology, as at the top of research interest, has potential to change our lives. Nanocomputers, nanocontrollers or nanomechanical devices will impact every aspect of marine technology. One of the nanotechnological products are nano-electromechanical systems (NEMS). NEMS are manufactured and/or assembled by a lot of nanoelements. One of possible elements is carbon nanotubes. Cold emission of electrons is the characteristic of nanotubes. When in contact with metallic object, such as ship, it is possible that undesirable electric flow occurs. So, this article contributes to calculation of electrostatic charge of a nanotube in the proximity of ships hall. The purpose of the research is to point out problems in application of nanotechnology on board

Influence of the Change in Relative Dielectric Constant on Partial Discharge in Insulators

This paper covers several topics relating to a single problem. The primary concern is partial discharge (PD). The second concern is the influence of environmental factors on PD, which is calculated by modifying the relative dielectric constant. The third concern is the manner of presentation of the obtained data. Finally, the basic concern is predicting what will happen to the PD when a parameter changes a material’s characteristics, what the margin of tolerance is and consequently, what will happen to a dielectric-dependent system under PD. The study of partial discharge in dielectrics is important for the reliability of electrical systems of vital importance aboard vehicles. This paper examines the effect of relative permittivity change to PD in dielectric materials. The topic was previously researched, but only for cables. PD modeling is carried out in Simulink using an equivalent electrical circuit. Simulation is carried out using the Matlab programming language using an m-script and Simulink model with parameter changes executed by script loop. Another topic of this paper is how to present the voluminous data obtained (in this case 3.5 GB), dealing with large data handling and visualization. 3D graphs with multi-curve representation in the same figure are used for visualization of the obtained results.