Abstract-A novel active contour model named VariableBackground Active Contour model is proposed and applied for the detection of thyroid nodules in ultrasound images. The new model offers edge independency, no need for smoothing, ability for topological changes and it is more accurate when compared to the Active Contour Without Edges model. Improved accuracy is achieved by introducing as background a limited image subset which appropriately changes shape to reduce the effects of background inhomogeneity. We validated the proposed model on ultrasound images acquired from 24 patients and the results demonstrate an improvement in accuracy when compared to the Active Contour Without Edges model.
Abstract. This paper presents a novel framework for thyroid ultrasound image segmentation that aims to accurately delineate thyroid nodules. This framework, named GA-VBAC incorporates a level set approach namedVariable Background Active Contour model (VBAC) that utilizes variable background regions, to reduce the effects of the intensity inhomogeneity in the thyroid ultrasound images. Moreover, a parameter tuning mechanism based on Genetic Algorithms (GA) has been considered to search for the optimal VBAC parameters automatically, without requiring technical skills. Experiments were conducted over a range of ultrasound images displaying thyroid nodules. The results show that the proposed GA-VBAC framework provides an efficient, effective and highly objective system for the delineation of thyroid nodules.
I. INTRODUCTIONLTRASOUND (US) is a widely used form of medical imaging, both as a primary modality and as an adjunct to other diagnostic procedures, providing substantial clues in differential diagnosis [1][2]. Its main advantages include non-invasiveness, low-cost and short acquisition times. The interpretation of US images is not a trivial task. A long learning curve is required for radiologists so as to acquire skills in recognizing the image features that comprise risk factors for different diseases, whereas it is difficult to remove the I. Legakis is with the Dept. of Endocrinology, Henry Dunant Hospital, Mesogion 107, 11526 Athens, Greece (e-mail: ilegak@med.uoa.gr).subjective element from the diagnostic process. A challenging task for US diagnostics is the assessment of thyroid nodules. Thyroid nodules are solid or cystic lumps formed in the thyroid gland. They may be caused by a variety of thyroid disorders and carry a considerable risk of malignancy. T he most useful features of the US images, which are usually correlated with the pathology of the thyroid nodules, are echogenicity, texture and shape [3]. For example, recent studies [3][4] indicate that nodules of irregular boundary are associated with a higher malignancy risk. A precise US image delineation method capable of capturing these features or indicating their presence to the experts could contribute to the objectification of medical decisions, and could also be used as an educational tool for trainee radiologists.Medical image segmentation approaches based on active contour models have been applied to images generated by medical imaging modalities as varied as US, magnetic resonance (MR), X -ray, computed tomography (CT) and angiography. Two-dimensional and three-dimensional active contours have been used to segment, visualize, track and quantify a variety of anatomic structures ranging in scale from the macroscopic to the microscopic. These structures include the heart, the cerebrum, a kidney, the lungs, objects such as brain tumors, a fetus, and even cellular structures such as neurons and chromosomes [5]. Cheng et al [6] developed and validated an automatic system using active contours for detecting the intimal and the adventitial layers so as to calculate the intima-media thickness of the common carotid artery. Plissiti et al [7] proposed an active contour model for the delineation of the lumen and media-adventitia border in sequential intravascular US frames. Chang et al [8] utilized a 3-D geodesic active contour to obtain the tumor contour for the pre-and the post-operative malignant breast excision by the vacuum assisted biopsy instrument Mammotome. Jeong et al [9] extended and combined the level set active contour segmentation approach and the agglomerative hierarchical k-means approach for unsupervised clustering. Their approach has been applied for the classification/differentiation of soft tissues in multiband high-resolution ultrasonic transmission tomography images.This wide applicability of active contours can be attributed to th...
In this paper, a novel computer-based approach is proposed for malignancy risk assessment of thyroid nodules in ultrasound images. The proposed approach is based on boundary features and is motivated by the correlation which has been addressed in medical literature between nodule boundary irregularity and malignancy risk. In addition, local echogenicity variance is utilized so as to incorporate information associated with local echogenicity distribution within nodule boundary neighborhood. Such information is valuable for the discrimination of high-risk nodules with blurred boundaries from medium risk nodules with regular boundaries. Analysis of variance is performed, indicating that each boundary feature under study provides statistically significant information for the discrimination of thyroid nodules in ultrasound images, in terms of malignancy risk. k-nearest neighbor and support vector machine classifiers are employed for the classification tasks, utilizing feature vectors derived from all combinations of features under study. The classification results are evaluated with the use of the receiver operating characteristic. It is derived that the proposed approach is capable of discriminating between medium-risk and high-risk nodules, obtaining an area under curve, which reaches 0.95. Keywords:Computer-Aided Diagnosis, Ultrasound, Thyroid Nodules, Boundary Features. IntroductionThe results of clinical research demonstrate that the presence of blurred or irregular thyroid nodule boundaries on ultrasound (US) images correlate with malignancy risk [1], [2]. In this light, the quantification of nodule boundary irregularity by boundary-based features could be valuable for malignancy risk assessment, contributing to the objectification of medical decisions. Such boundary-based features could be combined with intensity and textural information within an integrated computer-aided-diagnosis (CAD) tool.Previous attempts on CAD categorization of thyroid nodules on US images include evaluation of parameters from the gray level histogram of thyroid US images [3], [4], intensity features extracted by the utilization of Radon transform [5], textural features extracted from gray level spatial-dependence matrices [6], [7], and the application of discriminant
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.