BREAST FIBROADENOMA AUTOMATIC DETECTION USING K-MEANS BASED HYBRID SEGMENTATION METHOD

被引：0

作者：

Filipczuk, Pawel ^{[1
]}

Kowal, Marek ^{[1
]}

Obuchowicz, Andrzej ^{[1
]}

机构：

[1] Univ Zielona Gora, Inst Control & Computat Engn, Zielona Gora, Poland

来源：

2012 9TH IEEE INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI) | 2012年

关键词：

Image analysis; image segmentation; breast cancer; fibroadenoma; DIAGNOSIS;

D O I：

暂无

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

Fibroadenoma is a benign tumor that has some features similar to a malignant one. The aim of this study was to examine the impact of fibroadenoma cases on the results of the automatic breast cancer diagnostic system based on the quantitative morphometric analysis of fine needle biopsy microscopic images. The database of 50 patients (500 images) of benign and malignant lesions used previously in our research was enriched by an additional 25 patients (250 images) of fibroadenoma cases. Experiments were performed using the k-means based hybrid segmentation method. The system was tested on a set of real case medical images with promising results.

引用

页码：1623 / 1626

页数：4

共 20 条

[1] Improved Automatic Detection and Segmentation of Cell Nuclei in Histopathology Images
Al-Kofahi, Yousef
Lassoued, Wiem
Lee, William
Roysam, Badrinath
[J]. IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, 2010, 57 (04) : 841 - 852
[2] [Anonymous], 1987, INTRO BIOPSY INTERPR
[3] [Anonymous], 1992, DIGITAL IMAGE PROCES
[4] [Anonymous], INT J APPL MATH COMP
[5] [Anonymous], INT J APPL MATH COMP
[6] [Anonymous], METRIKA
[7] [Anonymous], 1997, MACHINE LEARNING, MCGRAW-HILL SCIENCE/ENGINEERING/MATH
[8] Automatic segmentation of overlapping nuclei with high background variation using robust estimation and flexible contour models
Clocksin, WF
[J]. 12TH INTERNATIONAL CONFERENCE ON IMAGE ANALYSIS AND PROCESSING, PROCEEDINGS, 2003, : 682 - 687
[9] Cloppet F, 2008, INT C PATT RECOG, P789
[10] Filipczuk P., 2010, J MED INFORM TECHNOL, V15, P193

← 1 2 →