|
Title:
|
NETWORK-BASED ENSEMBLE CLASSIFIER FOR DIFFERENTIAL DIAGNOSIS OF BREAST TUMORS IN ULTRASONIC IMAGES |
|
Author(s):
|
Atsushi Takemura |
|
ISBN:
|
978-972-8939-31-1 |
|
Editors:
|
Piet Kommers, Tomayess Issa and Pedro Isaías |
|
Year:
|
2010 |
|
Edition:
|
Single |
|
Keywords:
|
Ultrasonic image, breast tumor, differential diagnosis, network- based classifier, AdaBoost |
|
Type:
|
Poster/Demonstration |
|
First Page:
|
333 |
|
Last Page:
|
336 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
This paper proposes a new method for supporting differential diagnosis of breast tumors in ultrasonic image. The diagnostic support system is constructed by a network-based ensemble classifier trained by the AdaBoost machine learning algorithm with effective feature selection. The proposed system uses only an ensemble classifier from the remote database. In order to decrease the volume of transmission data and secure the system, the proposed system performed without original ultrasonic images in an image database. To improve the accuracy of automated differential diagnosis of breast tumors by using AdaBoost, the features based on statistics of K-distribution and Nakagami distribution were defined. Validation test using 310 carcinomas, 50 fibroadenomas, and 50 cysts showed the high performance of the proposed method of segmentation and discrimination. |
|
|
|
|
|
|
