Wyniki wyszukiwania - BazTech

1

Identification of Football Players based on Generic Fourier Descriptor Applied for the Recognition of Numbers

Frejlichowski D.

Image Processing & Communications

|

2016

|

Vol. 21, no. 4

13--18

EN

There are various possible approaches to the task of automatic identification of football players on video sequences. In the paper the numbers on jerseys are analysed and investigated. Firstly, the algorithm for the localisation of sport shirts and consecutively numbers is described. Secondly, the method for the recognition of the extracted numbers is proposed and experimentally investigated. The approach analysed in the paper applies the Generic Fourier Descriptor for the representation of extracted binary shapes of numbers.

2

Application of the Moment Shape Representations to the General Shape Analysis

Frejlichowski D.

Image Processing & Communications

|

2016

|

Vol. 21, no. 4

31--38

EN

The General Shape Analysis (GSA) is a task similar to the shape recognition and retrieval. However, in GSA an object usually does not belong to a template class, but can only be similar to some of them. Moreover, the number of applied templates is limited. Usually, ten most general shapes are used. Hence, the GSA consists in searching for the most universal information about them. This is useful when some general information has to be concluded, e.g. in coarse classification. In this paper the result of the application of three shape descriptors based on the moment theory to the GSA is presented. For this purpose the Moment Invariants, Contour Sequence Moments, and Zernike Moments were selected.

3

Application of Selected Geometrical 3D Object Description Algorithms to the Problem of Model Retrieval

Frejlichowski D.

Przegląd Elektrotechniczny

|

2015

|

R. 91, nr 2

103-106

EN

In the paper the comparison of three geometrical approaches to the representation of 3D shapes is presented. Two of them are wellknown and popular, i.e. Extended Gaussian Image and Shape Histograms. They are compared with a method based on the transformation of points from Cartesian to spherical co-ordinates. For the purpose of the experiments the „Princeton Shape Benchmark” database was applied, which became popular in the task of experimental evaluation of 3D shape descriptors. The current rapid development of computer hardware makes the processing of 3D scenes faster. Hence, the description of objects in image processing, recognition and indexing is possible. Amongst many existing applications of 3D shape descriptors their usage in CAD systems, biometrics, entertainment, virtual reality and image retrieval are especially tempting. In the paper the last listed problem is analysed.

PL

W artykule opisano porównanie trzech podejść geometrycznych do zagadnienia reprezentacji kształtu 3D. Dwa spośród nich to dobrze znane i popularne techniki, tj. Extended Gaussian Image oraz Shape Histograms. Zostały one porównane z metodą opartą na przekształceniu punktów z kartezjańskiego do sferycznego układu współrzędnych. Na potrzeby eksperymentu wykorzystano bazę „Princeton Shape Benchmark”, która stała się popularna w eksperymentalnej ocenie deskryptorów kształtu 3D. Obecny szybki rozwój sprzętu komputerowego przyspiesza przetwarzanie scen trójwymiarowych. Dzięki temu opis obiektów na potrzeby przetwarzania obrazów, rozpoznawania oraz indeksowania stał się możliwy. Pośród wielu istniejących zastosowań deskryptorów kształtu 3D ich użycie w systemach CAD, biometrii, rozrywce, wirtualnej rzeczywistości oraz indeksowaniu obrazów jest szczególnie pożądane. W niniejszym artykule ostatnie z wymienionych zastosowań jest analizowane.

4

A Study on Various Shape Descriptors and Matching Methods Applied in the General Shape Analysis

Frejlichowski D., Gościewska K.

Przegląd Elektrotechniczny

|

2015

|

R. 91, nr 2

89-92

EN

In the paper, the General Shape Analysis problem is investigated using various combinations of shape descriptors and matching methods. Five shape descriptors were used, namely the Two-Dimensional Fourier Descriptor, Generic Fourier Descriptor, UNL-Fourier, Zernike Moments and Point Distance Histogram, and four matching methods - the Euclidean distance, Mahalanobis distance, correlation coefficient and C1 metric. The experiments made it possible to determine how matching methods influence the final effectiveness when a particular shape descriptor was applied.

PL

W artykule przedstawiono badanie różnorodnych kombinacji deskryptorów kształtu i metod dopasowania w problemie Ogólnej Analizy Kształtu. Wykorzystano pięć deskryptorów, a mianowicie Dwuwymiarowy Deskryptor Fouriera, Generic Fourier Descriptor, UNL-Fourier, Momenty Zernike’a oraz Point Distance Histogram, a także cztery metody dopasowania – odległość Euklidesową, odległość Mahalanobisa, współczynnik korelacji oraz metrykę C1. Eksperymenty pozwoliły zdeterminować, jaki wpływ mają metody dopasowania na ostateczną skuteczność eksperymentu.

5

Intelligent video surveillance systems for public spaces – a survey

Zabłocki M., Gościewska K., Frejlichowski D., Hofman R.

Journal of Theoretical and Applied Computer Science

|

2014

|

Vol. 8, nr 4

13--27

EN

In recent years, a large number of cameras have been installed in public spaces as a part of intelligent video surveillance systems. Such systems are being continuously developed due to the advancements in the Video Content Analysis algorithms. In this paper, some of the latest state-of-the-art intelligent video surveillance systems will be presented in the context of their most desirable characteristics and features. Due to the variety of the solutions the following categories have been taken into consideration: systems based on object detection, tracking and movement analysis, systems able to warn against, detect and identify abnormal and alarming situations, systems based on vehicle detection and traffic or parking lots analysis, object counting systems, systems based on multiple integrated camera views, privacy preserving systems and systems based on cloud environment. The paper describes several solutions for each category and underlines main functionalities of the current intelligent surveillance systems.

6

The application of simple shape measures based on a Minimum Bounding Rectangle to the General Shape Analysis problem

Frejlichowski D., Gościewska K.

Journal of Theoretical and Applied Computer Science

|

2013

|

Vol. 7, nr 4

35--41

EN

This paper deals with the problem of General Shape Analysis. Simple shape measures based on a minimum bounding rectangle are investigated so as to obtain and compare general shape characteristics. The goal of General Shape Analysis is to find for each test object the one or few most similar templates. Thanks to this we can determine general information about a shape, such as how rectangular it is. Our work involves trying to find a shape descriptor that is simple to derive, allows for fast matching, and is the most consistent with a human benchmark that has been collected from the appropriate inquiry forms. Some experiments using various simple shape measures were performed. A concise explanation of these experiments and some percentage accuracy results are provided in this paper.

7

Application of the curvature scale space transform to the representation of three-dimensional models

Frejlichowski D.

Image Processing & Communications

|

2013

|

Vol. 18, no. 1

23--29

EN

In this paper an algorithm for the representation of 3D models is described and experimentally evaluated. Three-dimensional objects are becoming very popular recently and they are processed in various ways - analysed, retrieved, recognised, and so on. Moreover, they are employed in various aplications, such as virtual reality, entertainment, Internet, Computer Aided Design, or even in biometrics or medical imaging. That is why the development of appropriate algorithms for the representation of 3D objects is so important recently. These algorithms - so called 3D shape descriptors - are assumed to be invariant to particular transformations and deformations. One of the possible approaches is based on the projections of a 3D object into planar shapes and representation of them using a 2D shape descriptor. An algorithm realising this idea is described in this paper. Its first stage is based on the rendering of 20 2D projections, from various points of view. Later, the obtained projections are stored in a form of bitmaps and the Curvature Scale Space algorithm is applied for the description of the planar shapes extracted from them. The proposed approach is experimentally compared with several other 3D shape representation methods.

8

Detection of Erythrocyte Cells in Microscopy Images

Frejlichowski D.

Przegląd Elektrotechniczny

|

2012

|

R. 88, nr 10b

264-267

EN

In the paper an algorithm for automatic detection of red blood cells in microscopic images is described and investigated. For this purpose digital microscopic images stained by means of the MGG (May-Grunwald-Giemsa) method are applied. This method consists in three main stages. The first one is composed of the conversion to binary image using the modified histogram thresholding. The second one comprises the localisation and extraction of each separate object. The last stage covers the selection of erythrocytes and the rejection of other cells. The method described in the paper method gave the average efficiency above 83%, which is a promising result considering the specific characteristics of the investigated data as well as the large number of cells within an image.

PL

W artykule przedstawiono i omówiono algorytm przeznaczony do automatycznej detekcji czerwonych krwinek na obrazach mikroskopowych. Do tego celu użyto cyfrowych obrazów mikroskopowych, barwionych metodą MGG (May-Grunwald-Giemsa). Metoda opiera się na trzech głównych etapach. Pierwszy to konwersja do obrazu binarnego przy użyciu zmodyfikowanego progowania histogramowego. Drugi to lokalizacja i ekstrakcja każdego obiektu oddzielnie. Ostatni to selekcja erytrocytów i odrzucenie pozostałych komórek krwi. Opisywana w artykule metoda dała średnią skuteczność przekraczającą 83 %, co jest obiecującym wynikiem, biorąc pod uwagę specyficzne cechy badanych danych oraz dużą liczbę komórek na obrazie.

9

Application of the Curvature Scale Space Descriptor to the Problem of General Shape Analysis

Frejlichowski D.

Przegląd Elektrotechniczny

|

2012

|

R. 88, nr 10b

209-212

EN

In the paper the problem of General Shape Analysis (GSA) is investigated. It is similar to the recognition and retrieval of shapes. However, the analysed shape does not have to belong to any of the template classes. It is only similar to some of them. It may be said that the most general information about a shape is here concluded, namely how round, triangular, elliptical, etc. it is. This approach can be useful when an application uses only few general base classes. In the paper the Curvature Scale Space (CSS) shape description algorithm is applied to the problem.

PL

W artykule rozpatrywany jest problem Ogólnej Analizy Kształtu (OAK). Jest on zbliżony do rozpoznawania, a także indeksowania kształtów. Jednakże badany kształt nie musi należeć do żadnej z wzorcowych klas. Jest on jedynie podobny do niektórych z nich. Można powiedzieć, że w ten sposób wnioskujemy bardzo ogólną informację o kształcie – w jakim stopniu jest on okrągły, trójkątny, eliptyczny, itd. Omawiane podejście jest przydatne w zastosowaniach, w których stosuje się kilka podstawowych klas. W badaniach zastosowano deskryptor CSS.

10

SmartMonitor: recent progress in the development of an innovative visual surveillance system

Frejlichowski D., Gościewska K., Forczmański P., Nowosielski A., Hofman R.

Journal of Theoretical and Applied Computer Science

|

2012

|

Vol. 6, nr 3

28-35

EN

This paper describes recent improvements in developing SmartMonitor — an innovative security system based on existing traditional surveillance systems and video content analysis algorithms. The system is being developed to ensure the safety of people and assets within small areas. It is intended to work without the need for user supervision and to be widely customizable to meet an individual’s requirements. In this paper, the fundamental characteristics of the system are presented including a simplified representation of its modules. Methods and algorithms that have been investigated so far alongside those that could be employed in the future are described. In order to show the effectiveness of the methods and algorithms described, some experimental results are provided together with a concise explanation.

11

Analysis of possible system-level hardware implementation of the selected shape description algorithms

Frejlichowski D.

Journal of Theoretical and Applied Computer Science

|

2012

|

Vol. 6, nr 4

51-58

EN

This paper is devoted to the problem of the hardware implementation of selected shape description algorithms. The properties and complexity of particular methods are discussed in order to find those that are most applicable to a direct hardware implementation. Thus, the complexity of both the approaches and the processed data should be minimised. In this paper the properties of the Roundness, X/Y Feret, Point Distance Histogram, UNL, UNL-Fourier and Fourier Descriptors methods were investigated. Two aspects were analysed — the efficiency of an exemplary practical shape analysis application and the time consumption of the implemented algorithms.

12

Reprezentacja kształtu 3D z użyciem współrzędnych sferycznych

Frejlichowski D.

Metody Informatyki Stosowanej

|

2011

|

nr 3

41-47

EN

The paper presents experimental results on three dimensional shape description and recognition using a simple approach. The method uses the transformation from Cartesian coordinates of the vertices belonging to an object to the three-dimensional spherical ones. The described algorithm is tested using the objects deformed by affine transforms.

13

Aktualne trendy w tworzeniu systemów inteligentnego monitoringu wizyjnego

Frejlichowski D., Forczmański P., Nowosielski A., Hofman R.

Metody Informatyki Stosowanej

|

2011

|

nr 4

19-39

EN

This article provides an overview and critical analysis of computer vision algorithms used in the construction of the modules responsible for recognition and identification of objects which are elements of an intelligent monitoring system. We investigate state-of-the-art methods with an enough high potential to be implemented in a practical realization of such a system. The article describes three main elements of modern surveillance system, namely an adaptive background model, object extraction and tracking. Finally, we describe several recent benchmark datasets that can be used to test real systems.

14

Application of Zernike Moments to the problem of General Shape Analysis

Frejlichowski D.

Control and Cybernetics

|

2011

|

Vol. 40, no 2

515-526

EN

The Zernike Moments (ZM) have been successfully applied to the problem of shape recognition. Their properties allow for solving some fundamental problems in this task. Amongst them the most important one is the invariance to rotation, scaling, translation, and reflectional symmetry. Moreover, the obtained representation can vary according to the level of generalisation of a shape. For this reason in the paper the application of the ZM to the problem of General Shape Analysis (GSA) is proposed and experimentally investigated. The GSA problem is similar to the recognition and retrieval of shapes. However, only the most general classes of shapes (e.g. square, triangle, circle, ellipse) are assumed to perform the role of the basic templates. Moreover, the processed object does not have to belong to any of the template classes, but may be only similar to one of them. This enables us to receive the most general information about a shape, e.g. how square, triangular, round, elliptical, etc. it is. In the paper, in order to evaluate the Zernike Moments applied to the problem of GSA, the performance of this shape descriptor is compared with the results provided by nearly two hundred humans and collected by means of appropriate inquiry forms.

15

An experimental comparison of Extended Gaussian Image and Shape Distributions in 3D shape retrieval

Frejlichowski D.

Pomiary Automatyka Kontrola

|

2010

|

R. 56, nr 8

973-975

EN

The effective retrieval of three-dimensional shapes is a very crucial problem nowadays. It has to be not only efficient but also carried out in reasonable time. The last demand is especially difficult as 3D objects are usually built using lots of data (vertices, patches, etc.). That was the reason for minor interest dedicated few decades ago by scientists to them. At present, this problem became less important, thanks to the advances in computer hardware development. Now, one can find many new applications of 3D models, e.g. in CAD systems, entertainment, virtual reality, biometrics and image retrieval. In order to work with those objects three-dimensional shape descriptors are used. Those algorithms are created to represent objects independently of various problems concerning them, e.g. affine transformations, noise, occlusions. The result of experimental examination of two 3D shape descriptors is provided in the paper. The research was performed using the models from the "Princeton Shape Benchmark" database. This database is very popular in the task of experimental evaluation of 3D shape descriptors. In the paper two methods of that type are explored - Extended Gaussian Image and Shape Distributions - in the problem of 3D shape retrieval.

PL

Skuteczne wyszukiwanie kształtów trójwymiarowych w multimedialnych bazach danych jest istotnym problemem. Musi być ono nie tylko efektywne, ale i wykonywane w rozsądnym czasie. Ten drugi warunek jest szczególnie trudny do spełnienia, ponieważ obiekty 3D są zazwyczaj skonstruowane z użyciem dużej ilości danych (wierzchołki, powierzchnie, itp.). Było to dawniej powodem mniejszego zainteresowania naukowców tym zagadnieniem. Obecnie, problem ten stał się mniej znaczący, dzięki postępowi technicznemu w dziedzinie sprzętu komputerowego. Możemy więc aktualnie znaleźć wiele zastosowań modeli 3D, np. w komputerowo wspomaganym projektowaniu, rozrywce, rzeczywistości wirtualnej, biometrii oraz wyszukiwaniu obrazów. Aby móc pracować z tego typu obiektami stosowane są deskryptory kształtu. Te algorytmy są tworzone po to, by reprezentować obiekty niezależnie od poszczególnych problemów ich dotyczących, np. przekształceń afinicznych, szumu, okluzji. W artykule przedstawiono wyniki porównania eksperymentalnego dwóch deskryptorów kształtu 3D. Badania wykonano z użyciem modeli z bazy "Princeton Shape Benchmark". Baza ta jest bardzo popularna w ocenie deskryptorów kształtu 3D. W artykule dwie metody tego typu są badane - Rozszerzone Obrazy Gaussa oraz Rozkłady Kształtu - w kontekście problemu indeksowania kształtów 3D.

16

Badanie skutecznosci deskryptora Light Field w problemie reprezentacji kształtów trójwymiarowych na potrzeby ich indeksowania

Frejlichowski D.

Metody Informatyki Stosowanej

|

2010

|

nr 2 (23)

23-29

PL

W artykule przedstawiono rezultaty eksperymentu poświęconego określeniu wydajności deskryptora Light Field w problemie indeksowania kształtów trójwymiarowych. Metoda ta okazała się efektywna i uzyskała średni współczynnik skuteczności bliski 69 %. Wynik ten porównano z rezultatami deskryptora Extended Gaussian Image, który był testowany na tej samej bazie 312 modeli 3D (rezultaty tych badań opisano w [6]). Jak sie okazało algorytm LFD jest o blisko 9 procent skuteczniejszy w omawianym problemie. Uzyskany wynik należy uznać za stosunkowy dobry, wziąwszy pod uwagę duże podobieństwo pomiędzy sobą niektórych obiektów należących do różnych klas, przy jednoczesnej różnorodności obiektów wewnątrz danej klasy. Należy również podkreslić, że w eksperymentach pod uwagę brano tylko pierwszy wskazany w wyniku obiekt bazowy. O wiele lepsze rezultaty można byłoby osiagnąć, gdyby wybierać wiecej modeli, np. trzy lub pięć, co jest często spotykane w opisywanych w literaturze zadaniach indeksowania obiektów przechowywanych w dużych bazach multimedialnych.

EN

In the paper the results of the experimental evaluation of the Light Field Descriptor in the problem of three-dimensional shape representation are presented. The Light Field Descriptor (LFD) can be successfully used in the content-based three-dimensional model retrieval. The most general assumption in this algorithm is that if two 3D models are similar, they look similar from all viewing angles. Hence, in the explored approach several orthogonal projections of an object are rendered and used for its representation and later retrieval.

17

Analiza danych pomiarowych echosondy wielowiązkowej Simrad EM3000

Maleika W., Pałczyński M., Frejlichowski D., Stateczny A.

Metody Informatyki Stosowanej

|

2010

|

nr 4 (24)

55-64

EN

Errors of depth measurement are one of the most important problems in area of bathymetric data processing. The authors examined a data set aquired with the multibeam eehosounder Sim rad EM3000 in order to estimate the character of errors. For every measurement line formed by the survey points, the theoretical model of this fragment of bottom was defined using approximation. The differenees between this model and the survey points were treated as the measurement errors, which appeared to have a random distribution. The correlation of error value and angle between the beam and the vertical line could be observed. 99.7 % of er- ror values was smaller than 7 cm for depths between 10 and 20 m, while mean value of errors was approx. 3 cm.

18

Current challenges in content based image retrieval by means of low-level feature combining

Forczmański P., Frejlichowski D.

Annales Universitatis Mariae Curie-Skłodowska. Sectio AI, Informatica

|

2010

|

Vol. 10, no. 1

41--49

EN

The aim of this paper is to discuss a fusion of the two most popular low-level image features - colour and shape - in the aspect of content-based image retrieval. By combining them we can achieve much higher accuracy in various areas, e.g. pattern recognition, object representation, image retrieval. To achieve such a goal two general strategies (sequential and parallel) for joining elementary queries were proposed. Usually they are employed to construct a processing structure, where each image is being decomposed into regions, based on shapes with some characteristic properties - colour and its distribution. In the paper we provide an analysis of this proposition as well as the exemplary results of application in the Content Based Image Retrieval problem. The original contribution of the presented work is related to different fusions of several shape and colour descriptors (standard and non-standard ones) and joining them into parallel or sequential structures giving considerable improvements in content-based image retrieval. The novelty is based on the fact that many existing methods (even complex ones) work in single domain (shape or colour), while the proposed approach joins features from different areas.

19

Reprezentacja kształtu 3D oparta na deskryptorze EGI na potrzeby indeksowania

Frejlichowski D.

Metody Informatyki Stosowanej

|

2009

|

nr 3 (20)

83-89

PL

W niniejszym artykule przedstawiono wyniki wstępnych badań nad problemem opisu kształtów trójwymiarowych z użyciem deskryptora EGI (Extended Gaussian Image). Jest to jeden z popularniejszych algorytmów, którego największą zaletą jest intuicyjność i skuteczność w podstawowych zastosowaniach. W artykule skoncentrowano się na ozpoznawaniu kształtów w kontekście indeksowania tego typu obiektów w multimedialnych bazach danych (CBIR – Content Based Image Retrieval).

EN

The paper presents experimental results on three dimensional shape description and recognition using EGI (Extended Gaussian Image) 3D shape desciptor. The method uses the Gaussian image, which is obtained through mapping of areas and normal vectors for all polygons in an object into the Gaussian sphere. The descriptor is invariant to translation of an object in a scene. The algorithm will be applied to the problem of CBIR (Content Based Image Retrieval). More precisely, the 3D model retrieval is explored.

20

Analiza ogólnego kształtu obiektów wydobytych z obrazów cyfrowych rozpoznawanych z użyciem deskryptora PHD

Frejlichowski D.

Metody Informatyki Stosowanej

|

2009

|

nr 1 (18)

5-13

PL

W artykule zaprezentowano problem znajdowania ogólnego kształtu (np. kwadrat, koło, trójkąt) najbardziej podobnego do zadanego. Wyróżnia się on innym spojrzeniem na problem rozpoznawania kształtów. Dotąd zakładaliśmy poszukiwanie obiektu bazowego należącego do tej samej klasy co rozpoznawany. Jeżeli dla określonego obiektu nie udało się znaleźć odpowiedniego wzorca (np. gdy miara podobieństwa była zbyt niska) uznawaliśmy to za porażkę (wyjątkiem była sytuacja poprawnego odrzucenia wyników rozpoznawania, gdy odpowiedniej klasy po prostu nie ma w bazie). Tym razem z góry zakładamy, że badanego obiektu nie ma w bazie, a szukamy takiego wzorca, który jest do niego najbardziej podobny. Dzięki temu możemy określić jak bardzo ‘okrągły’ czy ‘kwadratowy’ jest dany kształt. Problem ten znaleźć może zastosowanie w wielu różnych dziedzinach, szczególnie na etapie wstępnej analizy kształtu, gdy najpierw chcemy go przydzielić do pewnej większej klasy, a dopiero potem w jej ramach szczegółowo badać.

EN

The paper presents experimental results of employing the Point Distance Histogram shape descriptor to the problem of finding the most similar general shape to the particular one. The template matching approach is used. However, unlike the typical recognition approach, here it is assumed that the test object is out of the template base. One of the general shapes, e.g. square, triangle, circle, ellipse, star, is indicated as the most similar to the test one. Three experiments were performed, using 50 various shapes.