Artūras Serackis

Spectrum Sensor Based on a Self-Organizing Feature Map

The paper focuses on the problem of burst signal detection in noisy radio environment. The ability to detect a burst signal in environment with low signal-to-noise level is crucial for cognitive radio solutions for which the detection of primary user transmission should be performed quickly and in energy efficient way. While the currently used spectrum sensing techniques based on the analysis of signal spectrum energy are simple to implement and are efficient, the sensitivity of these sensors in particular environment highly depends on the threshold estimation technique used in implementation. The classifier based on the self-organizing feature map proposed in this paper, takes the decision of the primary user signal presence in the measured environment. An experimental investigation was performed in 25 MHz wide frequency band on 949 MHz central frequency. The proposed spectrum sensing technique was compared with the alternatively proposed semi-adaptive threshold setting techniques for energy based spectrum sensors.

SOM based Clustering of Two-Dimensional Gel Image Segments for Quantitative Validation of Changes in Proteome

In this paper, we focus on the quantitative comparison of two-dimensional gel images in order to indicate the possible changes of the proteome in separate samples. The comparison of two or more separate gel images cannot be straightforward because it is hard to ensure the same experimental conditions for each experiment. The aim of our investigation was to increase the precision of quantitative validation of changes in proteome of analyzed biological tissue. Our work was divided into two parts: analysis of uncertainties during quantification of gel images prepared with different staining intensities, images with saturated protein spots and to propose an image processing technique to minimize these uncertainties. We proposed a SOM based clustering of watershed transformation based image segmentation result and separate comparison of protein spot intensities in each group. The experimental investigation showed, that the effective SOM size is 3×3 and background subtraction based on pixel intensities on image segment contour give less uncertainty comparing to alternative techniques.

Influence of Two Cameras Positioning to 3D Reconstruction

For last few years plenty of attention is dedicated to 3D object recognition and 3D reconstruction. Not a few produced robots perceive ambience with assistance of two cameras, analyzing perceived views. In this paper we analyze two camera positioning task. We change the angle between two cameras (principal axes) and analyze the influence of camera positioning to first steps of 3D reconstruction algorithm. One of the ways of camera mounting is in parallel, when all perceived points lie on the same plane. Unfortunately, because of influence of internal and external factors we receive different errors in 3D reconstruction. The influence of these errors can be reduced by minimizing object distance to camera and changing the angle between two cameras. Changing the angle between two cameras we change the object viewpoint. In our approach we show that even if object makes fairly wide angle with camera plane (60°–80°), we can estimate to 20 % of suitable feature points with selected Harris and Stephens corner detector. Ill. 5, bibl. 11 (in Lithuanian; summaries in Lithuanian, English, Russian).

Peer-to-peer adaptive video streaming system

The paper presents a peer-to-peer video streaming system with an ability to change the encoded video resolution adaptively to the data throughput changes. The low latency live video streaming based on the WebRTC specification was used in the investigation. The P2P video streaming system was able to predict the current performance of the data connection in the receiver and react by sending a request to the sender to increase or decrease the video resolution. An experimental investigation was performed by the use of network emulator and activating the 10% packet loss emulation. The system proposed in this paper was able to decrease the video resolution when the packet loss rate was increased. In addition, the system was able to detect the increase of the data throughput and to return the high video resolution for the video stream.

Heart beat prediction in distorted ECG signal using blood pressure measurements

The location of the heart beats in ECG signals are easily detectable by QRS detectors. However the quality of the ECG signal may vary considerably over time. There could be various sources of distortions that changes the shape of the ECG signal and makes some parts of the signal not readable. The QRS detector usually finds additional QRS complexes that are actually not present in the ECG signal. Paper presents effective algorithm to reduce the amount of false heart beat detections together with the algorithm to adjust heart beat position in noisy signal according to blood pressure measurements.

2DE gel image preprocessing using self-organising maps

Image processing for the comparison and matching of two images is usually based on classical cross-correlation technique. However matching of 2DE gel images requires more sophisticated approach. In this paper the preprocessing of 2DE gel images by self-organising map is proposed in order to improve results of image matching. Proposed technique divides 2DE gel image into regular image segments and extracts a set of segments, suitable for matching of two 2DE gels. The selection of image segments is made by classification accordingly to the statistical features of the image segments taken as the inputs of the self-organising map. Such preprocessing of 2DE gel image reduces number of segments for matching and selects a set of segments with smaller deviation of matching points for jointly used different distance metrics.