BazTech - Yadda

1

Differentiating age and sex in vertebral body CT scans – Texture analysis versus deep learning approach

Nurzynska Karolina, Piórkowski Adam, Strzelecki Michał, Kociołek Marcin, Banyś Robert Paweł, Obuchowicz Rafał

Biocybernetics and Biomedical Engineering

|

2024

|

Vol. 44, no. 1

20--30

EN

The automated analysis of computed tomography (CT) scans of vertebrae, for the purpose of determining an individual’s age and sex constitutes a vital area of research. Accurate assessment of bone age in children facilitates the monitoring of their growth and development. Moreover, the determination of both age and sex has significant relevance in various legal contexts involving human remains. We have built a dataset comprising CT scans of vertebral bodies from 166 patients of diverse genders, acquired during routine cardiac examinations. These images were rescaled to 8-bit data, and textural features were computed using the qMaZda software. The results were analysed employing conventional machine learning techniques and deep convolutional networks. The regression model, developed for the automatic estimation of bone age, accurately determined patients’ ages, with a mean absolute error of 3.14 years and R2 = 0.79. In the context of classifying patient gender through textural analysis supported by machine learning, we achieved an accuracy of 69 %. However, the application of deep convolutional networks for this task yielded a slightly lower accuracy of 59 %.

2

Classifying median nerves in carpal tunnel syndrome: Ultrasound image analysis

Obuchowicz Rafał, Kruszyńska Julia, Strzelecki Michał

Biocybernetics and Biomedical Engineering

|

2021

|

Vol. 41, no. 2

335--351

EN

Rationale and objectives: Carpal tunnel syndrome (CTS) refers to a common median nerve pathology, which is related to an increased pressure in the fibrous/bone canal of the wrist. Ultrasound gained popularity recently as a useful tool for the accurate and repetitive diagnosis of carpal tunnel syndrome. The present study aimed to develop an objective, repetitive technique for assessing median nerves based on carpal tunnel ultrasound texture analysis. Material and methods: Sixty ultrasound images, including 30 images of swollen ‘‘symptomatic” median nerves and 30 normal ‘‘asymptomatic” median nerves, were used in this study. Narrow age group of patients were selected. They were recruited after positive nerve conduction study and with present clinical symptoms reviled on basis of interview and written questionnaire. Meticulous nerve area and echogenicity assessment were conducted in line with existing recommendations. Results: Using the feature-selection tool MaZda, an exhaustive search of the data space was conducted, and four texture features were found for which the classification was the most accurate. Images were classified using a support vector machine with a five-fold cross-verification in MATLAB. Evoked outcomes showed a 79% correct classification rate. Conclusion: Computer analysis of the image echogenicity of the median nerve presented confidence levels comparable to trusted evaluation techniques. Further, it is a promising tool for assessing the nerve’s status in CTS as approach of the CTS assessment free from subjectivity of examiner. The developed method enables nerve classification based on echogenicity that reflects the nerve composition changes not only subjective nerve area assessment.

3

Prediction of forming limits and microstructural evolution during warm stretch forming of DP590 steel

Pandre Sandeep, Morchhale Ayush, Kotkunde Nitin, Singh Swadesh Kumar, Ravindran Sujith

Archives of Civil and Mechanical Engineering

|

2021

|

Vol. 21, no. 3

397--419

EN

Dual-phase (DP) steel has an excellent blend of various mechanical properties; hence it is used immensely in the automotive industries. It is challenging to form high strength DP steel into desirable complex shapes because of their limited formability at room temperature conditions. One of the proven alternatives is warm/hot forming. In-detail investigation of forming limits over DP590 steel has been carried out in present work. Firstly, various constitutive models and yield criteria have been formulated for DP steel at different temperatures and strain rates. The modified Arrhenius (m-A) constitutive model and Barlat 1989 yielding function displayed the best prediction of flow stress and anisotropic yielding behavior, respectively. The experimental forming limits (FLD) were evaluated at 300, 473 and 673 K temperatures using Nakazima tests. The forming limits of the material are improved by approximately 24% on increasing the temperature from 300 to 673 K. The textural analysis of the deformed surface has been done using electron back scattered diffraction (EBSD) studies, and γ fibers are found to be responsible for improvement in the formability of the material. Additionally, Marciniak and Kuczynski (MK) model was used to predict the theoretical FLD using all the possible combinations of constitutive models and yield criteria. Finally, the m-A constitutive model, along with Barlat 1989 yielding function has shown the best prediction for forming limits at all the temperatures. The finite element study has also been performed using mentioned material models for accurate prediction of dome height, surface strain and thickness distribution across the specimens.

4

The influence of the principal component analysis of texture features on the classification quality of sponge tissue images

Dzierżak Róża

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2020

|

T. 10, nr 3

13--16

EN

The aim of this article was to determine the effect of principal component analysis on the results of classification of spongy tissue images. Four hundred computed tomography images of the spine (L1 vertebra) were used for the analyses. The images were from fifty healthy patients and fifty patients diagnosed with osteoporosis. The obtained tissue image samples with a size of 50x50 pixels were subjected to texture analysis. As a result, feature descriptors based on a grey level histogram, gradient matrix, RL matrix, event matrix, autoregressive model and wavelet transform were obtained. The results obtained were ranked in importance from the most important to the least important. The first fifty features from the ranking were used for further experiments. The data were subjected to the principal component analysis, which resulted in a set of six new features. Subsequently, both sets (50 and 6 traits) were classified using five different methods: naive Bayesian classifier, multilayer perceptrons, Hoeffding Tree, 1-Nearest Neighbour and Random Forest. The best results were obtained for data on which principal components analysis was performed and classified using 1-Nearest Neighbour. Such an algorithm of procedure allowed to obtain a high value of TPR and PPV parameters, equal to 97.5%. In the case of other classifiers, the use of principal component analysis worsened the results by an average of 2%.

PL

Celem niniejszego artykułu było określenie wpływu analizy głównych składowych na wyniki klasyfikacji obrazów tkanki gąbczastej. Do analiz wykorzystano czterysta obrazów tomografii komputerowej kręgosłupa (kręg L1). Obrazy pochodziły od pięćdziesięciu zdrowych pacjentów oraz pięćdziesięciu pacjentów ze zdiagnozowaną osteoporozą. Uzyskane próbki obrazowe tkanki o wymiarze 50x50 pikseli poddano analizie tekstury. W wyniku tego otrzymano deskryptory cech oparte na histogramie poziomów szarości, macierzy gradientu, macierzy RL, macierzy zdarzeń, modelu autoregresji i transformacie falkowej. Otrzymane wyniki ustawiono w rankingu ważności od najistotniejszej do najmniej ważnej. Pięćdziesiąt pierwszych cech z rankingu wykorzystano do dalszych eksperymentów. Dane zostały poddane analizie głównych składowych wskutek czego uzyskano zbiór sześciu nowych cech. Następnie oba zbiory (50 i 6 cech) zostały poddane klasyfikacji przy użyciu pięciu różnych metod: naiwnego klasyfikatora Bayesa, wielowarstwowych perceptronów, Hoeffding Tree, 1-Nearest Neighbour and Random Forest. Najlepsze wyniki uzyskano dla danych, na których przeprowadzono analizę głównych składowych i poddano klasyfikacji za pomocą 1-Nearest Neighbour. Taki algorytm postępowania pozwolił na uzyskanie wysokiej wartości parametrów TPR oraz PPV, równych 97,5%. W przypadku pozostałych klasyfikatorów zastosowanie analizy głównych składowych pogorszyło wyniki średnio o 2%.

5

Comparison of the influence of standardization and normalization of data on the effectiveness of spongy tissue texture classification

Dzierżak Róża

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2019

|

T. 9, nr 3

66--69

EN

The aim of this article was to compare the influence of the data pre-processing methods – normalization and standardization – on the results of the classification of spongy tissue images. Four hundred CT images of the spine (L1 vertebra) were used for the analysis. The images were obtained from fifty healthy patients and fifty patients with diagnosed with osteoporosis. The samples of tissue (50×50 pixels) were subjected to a texture analysis to obtain descriptors of features based on a histogram of grey levels, gradient, run length matrix, co-occurrence matrix, autoregressive model and wavelet transform. The obtained results were set in the importance ranking (from the most important to the least important), and the first fifty features were used for further experiments. These data were normalized and standardized and then classified using five different methods: naive Bayes classifier, support vector machine, multilayer perceptrons, random forest and classification via regression. The best results were obtained for standardized data and classified by using multilayer perceptrons. This algorithm allowed for obtaining high accuracy of classification at the level of 94.25%.

PL

Celem niniejszego artykułu było porównanie wpływu metod wstępnego przetwarzania danych - normalizacji i standaryzacji - na wyniki klasyfikacji obrazów tkanki gąbczastej. Do analiz wykorzystano czterysta obrazów tomografii komputerowej kręgosłupa (kręg L1). Obrazy pochodziły od pięćdzisięciu zdrowych pacjentów oraz pięćdziesięciu pacjentów ze zdiagnozowaną osteoporozą. Uzyskane próbki tkanki (50×50 pikseli) poddano analizie tekstury w wyniku czego otrzymano deskryptory cech oparte na histogramie poziomów szarości, macierzy gradientu, macierzy RL, macierzy zdarzeń, modelu autoregresji i transformacie falkowej. Otrzymane wyniki ustawiono w rankingu ważności (od najistotniejszej do najmniej ważnej), a pięćdziesiąt pierwszych cech wykorzystano do dalszych eksperymentów. Dane zostały poddane normalizacji oraz standaryzacji, a następnie klasyfikowane przy użyciu pięciu różnych metod: naiwny klasyfikator Bayesa, maszyna wektorów wspierających, wielowarstwowe perceptrony, las losowy oraz klasyfikacji poprzez regresje. Najlepsze wyniki uzyskano dla danych na których przeprowadzono standaryzacje i poddano klasyfikacji za pomocą wielowarstwowych perceptronów. Taki algorytm postępowania pozwolił na uzyskanie wysokiej skuteczności klasyfikacji na poziomie 94,25%.

6

Texture-based identification of dystrophy phase. Indicating the most suitable features for therapy testing

Duda D.

Journal of Medical Informatics & Technologies

|

2018

|

Vol. 27

29--40

EN

In this study, texture analysis (TA) is applied for characterization of dystrophic muscles visualized on T2-weighted Magnetic Resonance (MR) images. The study proposes a strategy for indicating the textural features that are the most appropriate for testing the therapies of Duchenne muscular dystrophy (DMD). The strategy considers that muscle texture evolves not only along with the disease progression but also with the individual’s development. First, a Monte Carlo (MC) procedure is used to assess the relative importance of each feature in identifying the phases of growth in healthy controls. The features considered as age-dependent at a given acceptance threshold are excluded from further analyses. It is assumed that their application in therapies’ evaluation may entail an incorrect assessment of dystrophy response to treatment. Next, the remaining features are used in differentiation among dystrophy phases. At this step, an MC-based feature selection is applied to find an optimal subset of features. Experiments are repeated at several acceptance thresholds for age-dependent features. Different solutions are finally compared with two classifiers: Neural Network (NN) and Support Vector Machines (SVM). The study is based on the Golden Retriever Muscular Dystrophy (GRMD) model. In total, 39 features provided by 8 TA methods (statistical, filter- and model-based) are tested.

7

Comparative assessment of texture features for the identification of cancer in ultrasound images: a review

Faust O., Acharya U. R., Meiburger K. M., Molinari F., Koh J. E. W., Yeong C. H., Kongmebhol P., Ng K. H.

Biocybernetics and Biomedical Engineering

|

2018

|

Vol. 38, no. 2

275--296

EN

In this paper, we review the use of texture features for cancer detection in Ultrasound (US) images of breast, prostate, thyroid, ovaries and liver for Computer Aided Diagnosis (CAD) systems. This paper shows that texture features are a valuable tool to extract diagnostically relevant information from US images. This information helps practitioners to discriminate normal from abnormal tissues. A drawback of some classes of texture features comes from their sensitivity to both changes in image resolution and grayscale levels. These limitations pose a considerable challenge to CAD systems, because the information content of a specific texture feature depends on the US imaging system and its setup. Our review shows that single classes of texture features are insufficient, if considered alone, to create robust CAD systems, which can help to solve practical problems, such as cancer screening. Therefore, we recommend that the CAD system design involves testing a wide range of texture features along with features obtained with other image processing methods. Having such a compet-itive testing phase helps the designer to select the best feature combination for a particular problem. This approach will lead to practical US based cancer detection systems which deliver real benefits to patients by improving the diagnosis accuracy while reducing health care cost.

8

Analysis of the suitability of selected image types in a texture analysis of satellite imagery

Kupidura P., Staniak K.

Teledetekcja Środowiska

|

2017

|

T. 57

27--34

EN

The article presents studies on the impact of the source image type on the efficacy of image texture analysis in the terms of distinguishing classes of land use or land cover (LULC). Single gray-scale images are usually the inputs for this type of operation, however their selection is not unambiguous, especially in the case of multispectral images. Two very high resolution satellite images were used in the study: Pleiades (GSD: 2 m) and QuickBird (2.4 m). Five different input images were tested: the original near-infrared and red bands, the images of the first two main components, and the image of the normalised difference vegetation index - NDVI. Five LULC classes were compared to each other: bare soil, low vegetation, deciduous forests, coniferous forests and built-up areas. Granulometric analysis, as the one of the high efficient methods of texture analysis, was used for the test. Research results have shown that the choice of source image for this kind of processing can be very important for the efficacy of distinguishing between different LULC classes. NDVI images, and also the near infrared band and the first principal component were found most useful.

PL

Artykuł przedstawia badania dotyczące wpływu typu obrazu źródłowego na skuteczność analizy teksturowej obrazu z punktu widzenia wyodrębniania klas użytkowania lub pokrycia terenu (LULC). Tego typu operacjom poddawane są zazwyczaj pojedyncze obrazy w skali szarości, jednak ich wybór nie jest jednoznaczny, zwłaszcza w przypadku obrazów wielospektralnych. W badaniach wykorzystano dwa obrazy satelitarne o bardzo wysokiej rozdzielczości: Pleiades (GSD: 2 m) oraz QuickBird (2,4 m). Testowano pięć różnych obrazów wejściowych: oryginalne kanały bliskiej podczerwieni oraz czerwieni, obrazy dwóch pierwszych składowych głównych oraz obraz wskaźnika NDVI. Porównano wzajemnie pięć klas użytkowania lub pokrycia terenu: odkrytą glebę, niską roślinność, lasy liściaste, lasy iglaste oraz tereny zabudowane. Jako narzędzie testów wybrano analizę granulometryczną, jedną z metod analizy teksturowej o wysokiej skuteczności. Wyniki badań pokazały, że wybór obrazu źródłowego do przetworzeń może mieć bardzo duże znaczenie przy rozróżnianiu różnych klas użytkowania lub pokrycia terenu. Największą przydatnością cechowały się obrazy NDVI oraz kanału bliskiej podczerwieni i pierwszej składowej głównej.

9

Texture Analysis Method Based on Fractional Fourier Entropy and Fitness-scaling Adaptive Genetic Algorithm for Detecting Left-sided and Right-sided Sensorineural Hearing Loss

Wang S., Yang M., Li J., Wu X., Wang H., Liu B., Dong Z., Zhang Y.

Fundamenta Informaticae

|

2017

|

Vol. 151, nr 1/4

505--521

EN

To detect the sensorineural hearing loss (SNHL) from healthy people accurately, we used magnetic resonance imaging (MRI) to obtain the imaging data, and then proposed a new computer-aided diagnosis (CAD) system, on the basis of texture analysis method. In the first, we extracted 12-element feature from each brain image via fractional Fourier entropy (FRFE). Afterwards, multilayer perceptron (MLP) was employed as the classifier, which was trained by a novel fitness-scaling adaptive genetic algorithm (FSAGA). The statistical analysis over 49 subjects showed the overall accuracy of our method yielded 95.51%. Experimental results performed better than four state-of-the-art weight optimization methods, and this CAD system give significantly better performance than manual interpretation.

10

Żele spożywcze struktura i właściwości. Badania teksturalne, reologiczne i termiczne

Krystyjan M., Ciesielski W.

Przemysł Spożywczy

|

2016

|

T. 70, nr 2

13--17

PL

Żele spożywcze to najczęściej układy złożone, których właściwości zmieniają się w czasie oraz pod wpływem różnych czynników. Aby dokładnie śledzić zachodzące w nich zmiany, celowe wydaje się wykorzystanie najnowszych metod pomiarowych, jakich dostarcza nam nauka. W artykule przedstawiono możliwości wykorzystania technik teksturalnych, reologicznych oraz termicznych do badania struktury i właściwości żeli spożywczych. Omówiono m.in. test przebijania, TPA oraz test ekstruzji wstecznej i współbieżnej w badaniach teksturalnych. Scharakteryzowano pomiary dynamiczne, pełzania i powrotu oraz reksację naprężeń jako metody reologiczne. W przypadku pomiarów termicznych poruszono takie zagadnienia, jak termograwimetria, termiczna analiza różnicowa, skaningowa kalorymetria różnicowa oraz metody kombinowane.

EN

Food gels are, most frequently, the complex systems the properties of which change over the time and under the influence of different factors. To accurately trace the changes that take place in them, it is advisable to use the newest methods of measurement, which are provided by the science. The paper presents the possibilities of the application of textural, rheological and thermal techniques to study the structure and properties of food gels. There have been discussed the following problems: puncture test, TPA and retrograde and concurrent extrusion tests in the texture studies. Dynamics, creeping and return measurements and strain relaxation as rheological methods were characterized. In case of thermal measurements, such problems as thermogravimetry (thermogravimetric analysis), differential thermal analysis and differential scanning calorimetry were discussed.

11

Wykorzystanie granulometrii obrazowej w klasyfikacji treści zdjęć satelitarnych

Kupidura P.

Prace Naukowe Politechniki Warszawskiej. Geodezja

|

2015

|

z. 55

3--271

PL

W pracy przedstawiono nową metodę klasyfikacji treści zdjęć satelitarnych, opartą na wykorzystaniu granulometrycznej analizy tekstury obrazu. Opisano podstawy teoretyczne zaprezentowanej metody oraz zbadano jej dokładność, w zależności od wybranych parametrów przetworzeń granulometrycznych oraz cech obrazów źródłowych. Porównano ją także z innymi, dotychczas stosowanymi metodami klasyfikacji treści zdjęć satelitarnych. Istotą zaproponowanej metody jest wykorzystanie, oprócz danych spektralnych, również map granulometrycznych, czyli obrazów zawierających informację na temat tekstury obrazu w otoczeniu poszczególnych pikseli, powstających w wyniku granulometrycznych przetworzeń obrazu. Ważną zaletą granulometrii obrazowej jako metody oznaczania tekstury obrazu jest, m.in., wieloskalowość, czyli możliwość określania stopnia tekstury o rożnych rozmiarach ziarna. Drugą kluczową zaletą jest prawidłowe działanie również na krawędziach obiektów na obrazie, czyli odporność na tzw. błąd krawędzi. Przedstawiona metoda klasyfikacji polegająca na złożeniu map granulometrycznych i oryginalnych obrazów wielospektralnych pozwala uwzględniać kontekstową cechę interpretacyjną - teksturę, zwiększając możliwości klasyfikacji, a jednocześnie cechuje się dużą prostotą wykonania, podobną do klasycznej pikselowej klasyfikacji spektralnej. Efektywność granulometrii obrazowej zbadano pod kątem kilku czynników: rozdzielczości przestrzennej i rodzaju obrazu źródłowego, rodzaju morfologicznych operacji otwarcia i domknięcia oraz rozmiaru okna granulometrii określającego przestrzenny zasięg obliczenia lokalnej granulometrii względem poszczególnych pikseli. W pierwszej kolejności przeanalizowano separatywność wybranych klas pokrycia lub użytkowania terenu na podstawie wyłącznie danych spektralnych, a także na podstawie map granulometrycznych. W wybranych przypadkach, dzięki zastosowaniu analizy granulometrycznej, stwierdzono znaczny wzrost separatywności klas. Główna część pracy koncentruje się na badaniu dokładności klasyfikacji wykonanej przy użyciu zaproponowanej metody. Uzyskane wyniki dowodzą, że wykorzystanie map granulometrycznych w procesie klasyfikacji może znacząco podnieść jej dokładność. Stwierdzono przy tym istotny wpływ rozdzielczości obrazu źródłowego na efektywność badanej metody. Określono i opisano również znaczenie pozostałych, przedstawionych wyżej parametrów przetworzeń granulometrycznych, i samej klasyfikacji. Wnioski z badań pozwoliły na przedstawienie propozycji modelu dwuetapowej klasyfikacji wykorzystującej zarówno wyniki klasyfikacji spektralnej, jak i spektralno-teksturowej, co pozwoliło na uzyskanie optymalnej dokładności. Zaproponowana metoda może być stosowana w procesie półautomatycznego tworzenia map pokrycia lub użytkowania terenu na podstawie zdjęć satelitarnych lub lotniczych, pozwalając uzyskiwać większa dokładność, niż klasyfikacja w podejściu spektralnym.

EN

This book presents a new method of classification of satellite images, based on utilisation of granulometric analysis of image texture. The theoretical background of the method and its accuracy, depending on different parameters of granulometric processing and input images, is presented. It is compared to other approaches of satellite image classification. The essence qf the method relies on the use of granulometric maps, i.e. images containing information about a local texture in every pixel, additionally to spectral data contained in original multispectral images. One of the main advantages of the proposed method is its multiscality, i.e. a possibility to define a texture of an image, depending on a different size of texture element. Also, granulometric analysis of a texture is resistant to the so-called border error. As a result, it works properly on the edges of objects in an image. The method, based on a combination of granulometric maps and multispectral images. allows to take into account an important contextual feature of an image - that is, texture. Consequently, it is increasing a potential for correct classification, while remaining as simple as a pixel-based spectral classification approach. The effectiveness of image granulometry has been tested with different features and parameters: spatial resolution and a type of an input image, type of morphological opening and closing, as well as the size of a granulometric window, defining a range of a local granulometric analysis. A separability of different classes of land cover or land use, basing on spectral data and granulemetric maps, has been tested. Significant increase of separabillity has been observed in certain cases. The main goal of the book was to study accuracy of classification, basing on the presented method. The results of the research show that a use of granulometric maps in a classification process may increase the accuracy significantly. An important influence of input image's spatial resolution on the outcome has been observed. Also, the impact of other aforementioned features has been tested and described. Conclusions derived from the research allow to propose a two-step model, using results of both, spectral and spectro-textural classifications, to obtain an optimal accuracy of classification. The presented method may be used in process of semi-automatic generation of land cover and land use maps, basing on satellite or aerial images, obtaining accuracy level, which is higher than in the case of a spectral-based classification.

12

On the effect of image brightness and contrast nonuniformity on statistical texture parameters

Materka A., Strzelecki M.

Foundations of Computing and Decision Sciences

|

2015

|

Vol. 40, No. 3

163--185

EN

Computerized texture analysis characterizes spatial patterns of image intensity, which originate in the structure of tissues. However, a number of texture descriptors also depend on local average image intensity and/or contrast. This variations, known as image nonuniformity (inhomogeneity) artefacts often occur, e.g. in MRI. Their presence may lead to errors in tissue description. This unwanted effect is explained in this paper using statistical texture descriptors applied for MRI slices of a normal and fibrotic liver. To reduce the errors, correction of image spatial nonuniformity prior to texture analysis is performed. The issue of sensitivity of popular texture parameters to image nonuniformities is discussed. It is illustrated by classification examples of natural Brodatz textures, digitally modified to account for inhomogeneities – modeled as smooth variations of image intensity and contrast. A set of texture features is identified which represent certain immunity to image inhomogeneities.

13

Comparison of two interpolation methods for empirical mode decomposition based evaluation of radiographic femur bone images

Udhayakumar G., Sujatha C. M., Ramakrishnan S.

Acta of Bioengineering and Biomechanics

|

2013

|

Vol. 15, no. 2

73--80

EN

Analysis of bone strength in radiographic images is an important component of estimation of bone quality in diseases such as osteoporosis. Conventional radiographic femur bone images are used to analyze its architecture using bi-dimensional empirical mode decomposition method. Surface interpolation of local maxima and minima points of an image is a crucial part of bi-dimensional empirical mode decomposition method and the choice of appropriate interpolation depends on specific structure of the problem. In this work, two interpolation methods of bi-dimensional empirical mode decomposition are analyzed to characterize the trabecular femur bone architecture of radiographic images. The trabecular bone regions of normal and osteoporotic femur bone images (N = 40) recorded under standard condition are used for this study. The compressive and tensile strength regions of the images are delineated using pre-processing procedures. The delineated images are decomposed into their corresponding intrinsic mode functions using interpolation methods such as Radial basis function multiquadratic and hierarchical b-spline techniques. Results show that bi-dimensional empirical mode decomposition analyses using both interpolations are able to represent architectural variations of femur bone radiographic images. As the strength of the bone depends on architectural variation in addition to bone mass, this study seems to be clinically useful.

14

Skin region detection in digital images using discriminative textural features

Kawulok M.

Studia Informatica

|

2012

|

Vol. 33, nr 2B

37-48

EN

This paper presents a novel approach towards human skin regions detection and segmentation. The main contribution is concerned with proposing the discriminative texture analysis performed over skin probability maps obtained using conventional color-based methods. Results of the experimental validation reported in the paper confirm that the texture is an important source of information, neglected by many skin detection techniques.

PL

Niniejszy artykuł przedstawia nową metodę poprawy dokładności detekcji i segmentacji obszarów ludzkiej skóry w obrazach cyfrowych. Oryginalnym elementem przedstawionych badań jest zastosowanie dyskryminacyjnej analizy teksturalnej, przeprowadzanej w obrazie mapy prawdopodobieństwa występowania skóry. Mapy takie są otrzymywane za pomocą klasycznych metod funkcjonujących na bazie analizy barwy. Przedstawione wyniki walidacji eksperymentalnej potwierdzają, że tekstura stanowi istotne źródło informacji w detekcji skóry, co nie jest wykorzystywane przez większość istniejących metod.

15

Multistage morphological segmentation of bright-field and fluorescent microscopy images

Korzyńska A., Iwanowski M.

Opto-Electronics Review

|

2012

|

Vol. 20, No. 2

174-186

EN

This paper describes the multistage morphological segmentation method (MSMA) for microscopic cell images. The proposed method enables us to study the cell behaviour by using a sequence of two types of microscopic images: bright field images and/or fluorescent images. The proposed method is based on two types of information: the cell texture coming from the bright field images and intensity of light emission, done by fluorescent markers. The method is dedicated to the image sequences segmentation and it is based on mathematical morphology methods supported by other image processing techniques. The method allows for detecting cells in image independently from a degree of their flattening and from presenting structures which produce the texture. It makes use of some synergic information from the fluorescent light emission image as the support information. The MSMA method has been applied to images acquired during the experiments on neural stem cells as well as to artificial images. In order to validate the method, two types of errors have been considered: the error of cell area detection and the error of cell position using artificial images as the "gold standard".

16

New approach to prostate diagnosis - perfusion CT images analysis using "life belt" method

Śmietański J., Tadeusiewicz R.

Bio-Algorithms and Med-Systems

|

2012

|

Vol. 8, no. 1

145--158

EN

The most important task that could improve the efficacy of managing the prostate cancer (PCa) is to develop the technique which will be able to detect an existing PCa even in cases when currently used methods are insufficient. It is supposed that the perfusion computed tomography technology (p-CT) can improve the diagnosis of early PCa. Unfortunately, the perfusion prostate images are very difficult to analyze especially for doctors who are not enough experienced with such a kind of images. Therefore there is a need to find a computational method which could help the doctors to make the decision whether the prostate cancer exists or not and (if the results are positive) to correctly point out the cancerous region. In research which results are presented in the paper we analyzed a great number of prostate images derived from over 50 patients with proven or suspected PCa. We propose the new method, named “life-belt” which has significant potential for identifying cancerous regions.

17

Liver Tumour Classification Using Co-Occurrence Matrices on the Contourlet Domain

Kumar S. S., Moni R. S., Rajeesh J.

Machine Graphics and Vision

|

2011

|

Vol. 20, No. 2

197-214

EN

Liver disease is one of the most common diseases around the world, seriously affecting the health of humans. Computed tomography image based Computer Aided Diagnosis (CAD) could be crucially important in supporting liver cancer diagnosis. An effective approach to realize a CAD system for this purpose is described in this work. The CAD system employs automatic tumour segmentation, texture feature extraction and characterization into malignant and benign tumours. A Region of Inter- est (ROI) cropped from the automatically segmented tumour by confidence connected region growing and alternative fuzzy c means clustering is decomposed using multiresolution and multidirectional con- tourlet transform to obtain contourlet coefficients. Co-occurrence matrices of the contourlet coefficients are determined, and six parameters of texture characteristics, which include Angular Second Moment, Contrast, Correlation, Inverse Difference Moment, Entropy and Variance, are extracted from them. The extracted feature sets are classified into benign and malignant by a Generalized Regression Neural Net- work (GRNN) classifier. The performance of this scheme is evaluated by various performance measures and by the use a of the Receiver Operating Characteristic (ROC) curve. The results are compared with those obtained by a similar system using Wavelet Coefficients co-occurrence Matrix (WCCM) and Gray Level co-occurrence Matrix (GLCM) texture features. The results indicate that the proposed scheme based on the CCCM texture is effective for classifying malignant and begin liver tumours in abdominal CT imaging.

18

Use of Gabor filters for texture classification of airborne images and LiDAR data

Marmol U.

Archiwum Fotogrametrii, Kartografii i Teledetekcji

|

2011

|

Vol. 22

325--336

EN

In this paper, a texture approach is presented for building and vegetation extraction from LIDAR and aerial images. The texture is very important attribute in many image analysis or computer vision applications. The procedures developed for texture problem can be subdivided into four categories: structural approach, statistical approach, model based approach and filter based approach. In this paper, different definitions of texture are described, but complete emphasis is given on filter based methods. Examples of filtering methods are Fourier transform, Gabor and wavelet transforms. Here, Gabor filter is studied and its implementation for texture analysis is explored. This approach is inspired by a multi-channel filtering theory for processing visual information in the human visual system. This theory holds that visual system decomposes the image into a number of filtered images of a specified frequency, amplitude and orientation. The main objective of the article is to use Gabor filters for automatic urban object and tree detection. The first step is a definition of Gabor filter parameters: frequency, standard deviation and orientation. By varying these parameters, a filter bank is obtained that covers the frequency domain almost completely. These filters are used to aerial images and LIDAR data. The filtered images that possess a significant information about analyzed objects are selected, and the rest are discarded. Then, an energy measure is defined on the filtered images in order to compute different texture features. The Gabor features are used to image segmentation using thresholding. The tests were performed using set of images containing very different landscapes: urban area and vegetation of varying configurations, sizes and shapes of objects. The performed studies revealed that textural algorithms have the ability to detect buildings and trees. This article is the attempt to use texture methods also to LIDAR data, resampling into regular grid cells. The obtained preliminary results are interesting.

19

Quick texture generation for multiobject image analysis in brain pathology

Warowny M., Markiewicz T.

Przegląd Elektrotechniczny

|

2010

|

R. 86, nr 1

50-52

EN

The paper presents two methods of texture features generation for recognition between neoplasm and non-neoplasm cells in cancer diagnosis. There are few problems which need to be solved to achieve the best results: differentiable images, extraction of the individual cell image, selection of the most important features. We propose two models solving all of these problems. We compare the consequences of implementation Unser’s selected texture features and Markov Random Field model. The results of numerical experiments have shown in both methods quite good accuracy in recognizing cells. The proposed methods have proved to be useful in practical application at the diagnosis of cancer.

PL

Referat przedstawia zastosowanie generacji cech teksturalnych w rozpoznawaniu komórek nowotworowych. Proces rozróżniania komórek jest dość złożony ze względu naturalną złożoność obrazów, konieczność ekstrakcji pojedynczej komórki obrazu oraz trudności w wyborze odpowiednio różnicującej cechy. W pracy porównane zostały efekty zastosowania dwóch rodzajów modeli – opartego na cechach Unsera oraz modelu Markova. Główny nacisk pracy położony jest na praktyczne zastosowanie obu metod w diagnozie nowotworowej.

20

Description of biomedical textures by statistical properties of morphological spectra

Kulikowski J. L., Przytulska M., Wierzbicka D.

Biocybernetics and Biomedical Engineering

|

2010

|

Vol. 30, no. 3

19-34

EN

A class of mathematical model s of biological textures based on the multi-variable probability distributions of their morphological spectra is described. It is shown that a large class of such distributions can be presented by sufficient statistics consisting of the coefficients of their expansion into the series of multi-variable Hermite polynomials. The sufficient statistics can then be simplified by rejection of higher-order terms. The general concepts of mathematical models construction are illustrated by examples of textures of several biological tissues (aorta walls, liver and blood). The role of statistics based on absolute values of morphological spectral components and of their cross-correlation coefficients is underlined.