Melinda Kovacs

57218223954

Publications - 8

On Selecting, Ranking, and Quantifying Features for Building a Liver CT Diagnosis Aiding Computational Intelligence Method

Ferenc Lilik Melinda Kovacs Szilvia Nagy

Publication Name: Applied Sciences Switzerland

Publication Date: 2023-03-01

Volume: 13

Issue: 6

Page Range: Unknown

Description:

Featured Application: The selected attributes and their ranking and weights can be used in decision support algorithms in computer aided diagnosis. The liver is one of the most common locations for incidental findings during abdominal CT scans. There are multiple types of disease that can arise within the liver and many of them are nodular. The ultimate goal of our research is to develop an expert knowledge-based system using fuzzy signatures, to support decisions during diagnosis of the most frequent of these nodular lesions. Since the literature contains limited information about the graphical properties of CT images that must be taken into consideration and their relationship to one another, in this paper we focused on selecting and ranking the input parameters using expert knowledge and determining their importance. Six visual attributes of lesions (size, shape, density, homogeneity contour, and other features) were selected based on textbooks of radiology and expert opinion. The importance of these attributes was ranked by radiologist experts using questionnaires and a pairwise comparison technique. The most important feature was found to be the density of the lesion on the various CT phases, and the least important was the size, the order of the other attributes was other features, contour, homogeneity, and shape, with a Kendall concordance coefficient of 0.612. Weights for the attributes, to be used in the future fuzzy signatures, were also determined. As a last step, several statistical parameter-based quantities were generated to represent the above abstract attributes and evaluated by comparing them to expert opinions.

Open Access: Yes

DOI: 10.3390/app13063462

On Applying Expert Knowledge for Spline-Based Segmentation of Liver on Computed Tomography Images

Melinda Kovacs Patrik Gabor Rozsa Szilvia Nagy

Publication Name: 2023 58th International Scientific Conference on Information Communication and Energy Systems and Technologies Icest 2023 Proceedings

Publication Date: 2023-01-01

Volume: Unknown

Issue: Unknown

Page Range: 21-24

Description:

The segmentation of the liver from the surrounding soft tissues between the ribs on a non-enhanced, arterial or delayed phase computed tomography image complicated, as the muscle and many other soft tissues have very similar X-ray attenuation to the liver, thus they manifest as similar intensity and similar pattern domains on the computed tomography images. However, if the anatomical setup of the human body and the location of the bones, that have much higher attenuation are taken into account, a sufficiently smooth and precise borderline can be drawn between the ribcage and the liver, that can be developed into a liver contour by successive erosion and active contour methods.

Open Access: Yes

DOI: 10.1109/ICEST58410.2023.10187395

On the Applicability of Fuzzy Lines in Circular Hough Transform in Lesion Segmentation on Liver CT Images

Melinda Kovacs Szilvia Nagy

Publication Name: Studies in Computational Intelligence

Publication Date: 2023-01-01

Volume: 1040

Issue: Unknown

Page Range: 45-54

Description:

Most of the lesions that grow in the liver and need to be found on a CT take are roundish, though especially in the case of malignant lesions the irregularity of the shape is also rather common. Classical Hough transform, however, which is one of the most used methods for finding circles, usually fails to find the contours of these objects, because of the larger or smaller irregularity of the shapes. Introducing a fuzzification in the edge filtered version of the image, which is usually the basis of a Hough transform, makes the Hough transform more flexible for using in liver CT image analysis.

Open Access: Yes

DOI: 10.1007/978-3-031-07707-4_6

On the aggregation functions used in fuzzy signatures based medical image analysis

Ferenc Lilik Melinda Kovacs L. T. Koczy Szilvia Nagy

Publication Name: IEEE 23rd International Symposium on Computational Intelligence and Informatics Cinti 2023 Proceedings

Publication Date: 2023-01-01

Volume: Unknown

Issue: Unknown

Page Range: 409-414

Description:

The paper proposes the use of fuzzy signatures for modeling and analysis of pre-processed medical images, as an example, CT images of the liver are analyzed. Fuzzy signatures are used for the case of distinguishing larger and smaller malignant lesions from each other and from other (benign) nodular diseases in liver computed tomography images. As computed tomography phases are sometimes missing, the treatment of missing data is also briefly addressed. As the size of the malignant lesion influences its manifestation on the images, separate sub-signatures are developed for large and small lesions with the size being a separate layer of the signature. From the medical experts' point of view besides the tree structure of the signature it is crutial to determine the aggregations themselves, which model the ways experts fuse and combine the available information. For the subtrees for small and large lesions in the sub-roots algebraic multiplication seems to be the best fitting t-norm, while in the subtree weighted means.

Open Access: Yes

DOI: 10.1109/CINTI59972.2023.10381986

On using fuzzy c-means clustering in the fuzzy signature concept classification of liver lesions

Ferenc Lilik Melinda Kovacs L. T. Koczy Szilvia Nagy

Publication Name: International Conference on Electrical Computer Communications and Mechatronics Engineering Iceccme 2022

Publication Date: 2022-01-01

Volume: Unknown

Issue: Unknown

Page Range: Unknown

Description:

Liver is a very unique organ, it has double blood supply, not only through the arteries, but also through the veins. This property makes the contrast material enhanced computer tomography images show very characteristic behavior, depending on the time passed from the adjustment of the contrast material. When diagnosing a nodule in the liver by computer tomography, radiologist experts use multiple images with different delay factors, and generally, five basic characteristic properties of the nodule compared to the normal liver tissues. In the following considerations, we give a simplified model that reproduces the way medical experts take decisions, and offer a possibility to develop a computer aided diagnosis method. The classification of the nodules applies a model with fuzzy signatures, where the aggregation functions in the intermediate nodes are representing the radiologist point of view, while the membership degrees/functions at the leaves of the fuzzy signature's rooted tree are obtained from calculations applying the fuzzy c-means clustering algorithm.

Open Access: Yes

DOI: 10.1109/ICECCME55909.2022.9988684

Interpolative decisions in the fuzzy signature based image classification for liver CT

Ferenc Lilik Melinda Kovacs Szonja Krisztina Szujo L. T. Koczy Szilvia Nagy

Publication Name: IEEE International Conference on Fuzzy Systems

Publication Date: 2021-07-11

Volume: 2021-July

Issue: Unknown

Page Range: Unknown

Description:

In computer aided diagnostics image processing and classification plays an essential role. Image processing experts have been developing solutions for different types of problems, that can be related to image processing, however, due to the sensitivity of the data and the high cost of medical experts, these experimental methods usually have very limited use in real medical practice. The databases that are available are very limited, thus the elsewhere usual and extremely effective convolutional neural network or other automated learning methods are not so easy to adjust for medical image processing. To overcome this difficulty, this paper presents an expert knowledge based method which describes the decision structures by fuzzy signatures. Values of various properties of Computer Tomography images as e.g. density or homogeneity are being considered in these signatures that are different in all case of liver diseases. Because of the low number of samples available, fuzzy sets that describes the leafs of the signatures leads to sparse systems, hence interpolation is needed. However further investigations of other interpolation methods are planned, Stabilized Koczy-Hirota interpolation seems to be appropriate.

Open Access: Yes

DOI: 10.1109/FUZZ45933.2021.9494401

On the Applicability of Fuzzy Rule Interpolation and Wavelet Analysis in Colorectal Image Segment Classification

Ferenc Lilik Brigita Sziová Melinda Kovacs L. T. Koczy Szilvia Nagy

Publication Name: Studies in Fuzziness and Soft Computing

Publication Date: 2021-01-01

Volume: 394

Issue: Unknown

Page Range: 243-255

Description:

The automatic detection of colorectal polyps could serve as a visual aid for gastroenterologists when screening the population for colorectal cancer. A fuzzy inference based method was developed for determining whether a segment of an image has polyps. Its antecedent dimensions were the mean pixel intensity, the intensity’s standard deviation, the edge density, the structural entropies and the gradients, not only for the original image segments, but for its wavelet transformed versions. The method performed moderately well, even though the number of the input parameters was very large. In the present contribution we studied, that based on the necessary and usually applied conditions of the applicability of fuzzy rule interpolation, which antecedent dimensions should remain, and how omitting the other input parameters influences the results of the method.

Open Access: Yes

DOI: 10.1007/978-3-030-54341-9_21

Fuzzy Hough transformation in aiding computer tomography based liver diagnosis

Brigita Sziová Melinda Kovacs L. T. Koczy Szilvia Nagy

Publication Name: IEEE AFRICON Conference

Publication Date: 2019-09-01

Volume: 2019-September

Issue: Unknown

Page Range: Unknown

Description:

In the liver many types of roundish lesions can appear, as well as near the liver. Finding the contour of such objects can improve both the segmentation of the liver from its environment, and the segmentation of the lesions within the liver. However, classical Hough transform, which is one of the main methods for finding objects described by a predefined parameterized formula, usually fails to identify these object as they possess not perfectly round or elliptic contours. A fuzzification of the Hough transform is described and suggested for using in image preprocessing for liver diagnosis based on CT images in this paper. Fuzzifying the Hough transform improves the detection of roundish contours.

Open Access: Yes

DOI: 10.1109/AFRICON46755.2019.9133793