A novel automatic tumor detection for breast cancer ultrasound Images

Abstract

Breast cancer is one of the leading cause of death among women. Because of harmless and low cost, ultrasound is one of the most often used methods for breast cancer detection. However, tumor detection is very difficult in ultrasound images due to their specular nature and low quality of ultrasound images and most of the existing methods need to manually select ROI. In this paper, we proposed a novel automatic method for breast ultrasound (BUS) image tumor detection. We using the fuzzy logic theory and transform an US image into fuzzy domain. An iterative method is used to find threshold. The experimental results demonstrate that the proposed approach can automatically detect the tumor in BUS image with high accuracy. It can process low quality ultrasound image very well.

Detecção De Massas Em Imagens Da Mama Usando Índices De Diversidade E Algoritmos De Segmentação Em Grafo

“…Em Giuliato, Barcellos e Dias, (2008) é utilizado um filtro de suavização anisotrópica via Equações Diferenciais Parciais na fase de pré-processamento para a detecção de regiões suspeitas em mamografias. Em Zhang (2011) é proposto um algoritmo automático para detecção de tumores em imagens mamográficas através do uso da lógica de fuzzy. Este artigo propõe uma nova metodologia capaz de dar suporte ao diagnóstico médico na tarefa de detecção de anomalias na mama através do uso de imagem médica da mama.…”

A Two-Step Segmentation Method for Breast Ultrasound Masses Based on Multi-resolution Analysis

“…However, BUS images are typically characterized by speckle noise, shadows or other artifacts and poor edge definition, which are intrinsic to the imaging acquisition process and may result in a difficult and subjective analysis, even for experienced radiologists and oncologists ( Fig. 1) (Noble and Boukerroui 2006;Sehgal et al 2006;Zhang et al 2011). Therefore, computer-aided diagnosis (CADx) systems may become useful for both radiologists and oncologists.…”

“…It is frequently used as a follow-up technique or as an adjunct to mammography in detection and diagnosis. Although mammography is currently the most widely used imaging method, breast ultrasound (BUS) imaging has been emphasized as a valuable tool for early cancer detection and diagnosis because of its attractive properties (Zhang et al 2011). However, BUS images are typically characterized by speckle noise, shadows or other artifacts and poor edge definition, which are intrinsic to the imaging acquisition process and may result in a difficult and subjective analysis, even for experienced radiologists and oncologists ( Fig.…”

“…Morphologic operations were first applied to obtain more accurate contours and prevent typical watershed oversegmentation (Sarpe 2010). The watershed transform was also used in Zhang et al (2011) for automatic lesion detection. Images were preprocessed by mean filtering and fuzzy logic histogram thresholding.…”

Enlargement of Efficient Automatic Volumetric Mammographic Image Classification Based on Robust Features.

“…Breast region is exacted from the decimated image which is divided into three partitioned regions, the fat, the glandular and the dense region [6]. The optimal set of features selected by the Genetic algorithm is fed as input to Adaptive Neuro-fuzzy inference system for classification of images into normal, suspect and abnormal categories [7].Breast cancer ultrasound images are approached to detect the breast cancer based on the fuzzy logic theory, transform into a fuzzy domain and focuses on discovery a reliable ROI instead of finding the precise tumor place at ROI generation phase [8,9].The parameters for shape and margin of masses10 such as lifetime risk, percentage density are computed which shows the experimental findings are matched with the breast image reporting and data system standardization. In a case of histogram intersection method, the textual shape of the dynamic region and quantitative measurements in [11,12].All these methods discussed above have the problem of classifying false positive classes of mammographic images.…”