An algorithm for automatic ROI generation for detection of breast tumors in ultrasound images

سال انتشار: 1396
نوع سند: مقاله کنفرانسی
زبان: انگلیسی
مشاهده: 510

نسخه کامل این مقاله ارائه نشده است و در دسترس نمی باشد

این مقاله در بخشهای موضوعی زیر دسته بندی شده است:

استخراج به نرم افزارهای پژوهشی:

لینک ثابت به این مقاله:

شناسه ملی سند علمی:

ICBCMED13_065

تاریخ نمایه سازی: 2 تیر 1397

چکیده مقاله:

Introduction:Discovering early symptoms of breast cancer using clinical examination play a crucial role inmortality reduction. Among the available imaging modality breast ultrasound (BUS) is a key for early detection of breast malignancies. However, clinical experience and expert knowledge are necessary to achieve correct diagnosis. Recently breast ultrasound (BUS) Computer-aided diagnosis systems(CAD) systems have been introduces to overcome drawback of human perception-based diagnosis.Aims:One of the challenging problems for developing a CAD for BUS images islocating regions of interest (ROIs)automatically. This is because of complicated structure of breast and poor quality of ultrasound images.In this work, we propose anapproach for performing accurate and robust ROI generation.Material & Methods:ROI generation was mainly based on image entropy and gradient. It includes three main steps first, preprocessing to speckle reduction, contrast enhancement and the removal of non-tumor edge using appropriate spatial filters second, calculating textural informationentropy images and finally directional gradient. Result&Discussion:The proposed fully automatic segmentation method is applied to a BUS image and the performance is evaluated by the area and boundary error metrics. Compared with the manual ROI detection, the proposed method has comparable accuracyin segmenting BUS images.Conclusion:From the study result,it can be concluded that this algorithm is applicable in CAD based diagnostic of BUS images. Future work focuses on evaluating this algorithm on large set of database especially in more noisy BUS images with potentially multiple ROI numbers.

کلیدواژه ها:

breast tumor – ultrasound –image processing -ROI

نویسندگان

Reza Reiazi

PhD Assistance Professor of Medical Physics, Department of Medical Physics, School of Medicine, Iran University of Medical Science, Tehran, Iran

Reza Paydar

PhD Assistance Professor of Medical Physics, Department of radiation sciences, School of Allied Medicine, Iran University of Medical Science, Tehran, Iran

Hamed Bagheri

MSc Radiation and Wave research center, AJA University of medical sciences,

Maryam Etedadialiabadi

MS Radiology Department, Milad Hospital, Tehran, Iran