Martin Zukal

Automated localization of Temporomandibular Joint Disc in MRI images

This paper deals with localization of Temporomandibular Joint Disc (TJD) in Magnetic Resonance Images (MRI). Since the contrast of the TJD is quite low when compared to noise ratio when displayed using MRI, its detection is quite complicated. Therefore the method described in this paper are not not focused the disk itself but detect the most significant objects around TJD, which has usually much higher contrast. For the automatic TJD localization asessment, a training set containing 160 training samples (80 positive and 80 negative) were created and published and several approaches were examined to find the best method. The best results were achieved using support vector machine with Gaussian kernel, which achieved 98.16±2.81% accuracy of detection. The creation of the training models for feature extraction and model evaluation was implemented with RapidMiner tool and the IMMI extension. The models created are published at the IMMI extension homepage and they can also serve as a guide to use of the IMMI extension.

Evaluation of interest point detectors for scenes with changing lightening conditions

The paper is aimed at the description of different image interest point detectors and their properties. Particularly, the Harris-Laplace detector, the Fast Hessian detector and the Difference of Gaussian detector are described. These algorithms have already been evaluated with respect to common geometrical transformations such as the rotation, the scale change, etc. This paper describes the testing process and the impact of brightness change and histogram equalization on the repeatability of tested detectors. The evaluation has been performed on two different image databases containing altogether four hundred and eighty nine images. The repeatability has been used for the evaluation of the described interest point detectors. The best results have been achieved for the Fast Hessian detector which has proved to be the fastest and also the most robust. The repeatability of the Fast Hessian detector has reached the value of 65.39% after performing the histogram equalization on the Caltech database.

Vehicle license plate detection and recognition using symbol analysis

License plate detection and recognition is one of the most important aspects of applying computer techniques towards intelligent transportation systems. Detecting the accurate location of a license plate from a vehicle image is the most crucial step of a license plate detection system. This paper describes a proposing of a new region-based license plate detection method based on a symbol analysis. Throughout the detection and recognition the original image is filtered, transformed to gray-scale image and thresholded. In the next step the best candidates of regions are selected. The whole system was tested on fifty different cars with various license plates. The indication rate of success recognition is eighty eight percent.

Efficiency of BCH codes in digital image watermarking

Nowadays, the copyright laws protecting the digital content are getting quite common. Protecting the original or compressed images against illegal usage can be realized by means of watermarking methods. Nevertheless, various types of attacks are still a problem for many watermarking algorithms. In this paper, we proposed watermarking algorithms for grayscale and color images resisting to many common attacks. The discrete cosine transform (DCT), and block based forward error correction codes (FEC) are used for watermark embedding and extraction. Testing with different kinds of grayscale images demonstrates the robustness of the proposed algorithms to a lot of common attacks and effect of FEC.

Secure authentication and key establishment in the SIP architecture

This article deals with security features of SIP and focuses on advanced authentication and key establishment methods. These methods are described and a proposal of comprehensive security for the SIP architecture is presented in this article. The solution provides strong secure authentication of SIP users and protects against various attacks occurring in the SIP architecture. Furthermore, our proposal is implemented and tested. The experimental results from our proof of concept implementation are presented at the end of the text. Moreover, the conventional HTTP Digest method is also implemented and its performance is compared with methods in the proposal.

Analysis of Carotid Artery Transverse Sections in Long Ultrasound Video Sequences

Examination of the common carotid artery (CCA) based on an ultrasound video sequence is an effective method for detecting cardiovascular diseases. Here, we propose a video processing method for the automated geometric analysis of CCA transverse sections. By explicitly compensating the parasitic phenomena of global movement and feature drift, our method enables a reliable and accurate estimation of the movement of the arterial wall based on ultrasound sequences of arbitrary length and in situations where state-of-the-art methods fail or are very inaccurate. The method uses a modified Viola-Jones detector and the Hough transform to localize the artery in the image. Then it identifies dominant scatterers, also known as interest points (IPs), whose positions are tracked by means of the pyramidal Lucas-Kanade method. Robustness to global movement and feature drift is achieved by a detection of global movement and subsequent IP re-initialization, as well as an adaptive removal and addition of IPs. The performance of the proposed method is evaluated using simulated and real ultrasound video sequences. Using the Harris detector for IP detection, we obtained an overall root-mean-square error, averaged over all the simulated sequences, of 2.16 ± 1.18 px. The computational complexity of our method is compatible with real-time operation; the runtime is about 30-70 ms/frame for sequences with a spatial resolution of up to 490 × 490 px. We expect that in future clinical practice, our method will be instrumental for non-invasive early-stage diagnosis of atherosclerosis and other cardiovascular diseases.

Corner detectors: Evaluation of information content

This paper aims at the comparison of different corner detectors in terms of information content. Specifically, the FAST, Harris, Harris-Laplace and KLT detector are tested. We prove a hypothesis that the neighborhood of points detected by interest point detectors conveys more information than randomly selected points. The information content was computed from the 8x8 neighborhood of the detected points using entropy. From the qualitative point of view, we have observed that the Harris detector performed best in this measurement and was closely followed by the KLT detector.

Interest points as a focus measure

In this paper, we propose a novel focus measure that is based on algorithms for interest point detection, particularly on the Fast Hessian detector. The proposed measure is compared with the energy of image gradient and sum-modified Laplacian that are commonly used as focus measures to test its reliability and performance. The measures have been tested on a newly created database containing 84 images (12 images for seven objects). Our algorithm proved to be a good focus measure satisfying all the requirements described in the paper, in some cases it outperformed the other two measures.

Robustness evaluation of corner detectors for use in ultrasound image processing

This article deals with evaluation of suitability of interest point detectors (specifically corner detectors) for utilization in ultrasound image processing. Namely, the Harris detector, FAST (Features from Accelerated Segment Test) detector and Kanade-Lucas-Tomasi (KLT) detector have been tested on two data sets of ultrasound (US) images. The evaluation process consists of three experiments in which the images have been artificially corrupted by different types of noise or their brightness has been uniformly changed. The FAST detector proved to be more robust against noise than Harris and KLT detectors whereas the Harris and KLT detector outperformed FAST detector in case of brightness change.

Tracking and speed estimation of selected object in video sequence

Object tracking and speed estimation is one of the important tasks in video processing. The availability of high quality and cheap video camcorders and the increasing need for automated video analysis has generated a great deal of interest in video processing algorithms. Numerous approaches for speed estimation have been proposed. This paper deals with object tracking and speed estimation in a video sequence. We proposed a novel method to estimate the speed of moving object in video sequences captured with a static camcorder. The movement of object is determined by direction and speed. The article describes the algorithms used in the novel technique that were implemented in Rapidminer. One new operator for speed measurement was developed and implemented. This operator provides information about the movement of selected object in video sequence. Finally, three short shots were captured to test of the proposed solution. Results of these tests are presented at the end of this paper.