Àá½Ã¸¸ ±â´Ù·Á ÁÖ¼¼¿ä. ·ÎµùÁßÀÔ´Ï´Ù.
KMID : 1144120120020020100
Biomedical Engineering Letters
2012 Volume.2 No. 2 p.100 ~ p.110
Lymphocyte image segmentation using functional link neural architecture for acute leukemia detection
Mohapatra Subrajeet

Patra Dipti
Kumar Sunil
Satpathy Sanghamitra
Abstract
Purpose: Microscopic examination of stained blood slides is an indispensable technique for hematological disease recognition. Diagnosis based on human visual interpretation is often subjected to inter and intra observer variations, slowness, tiredness and operator experience. Accurate and authentic diagnosis of hematological neoplasia is essential in the planning of suitable surgery and chemotherapy. This paper aims at proposing a fast and simple framework for lymphocyte image segmentation.

Methods: Accurate segmentation of lymphocyte is essential as it facilitates automated leukemia detection in blood microscopic images. In this paper image segmentation is considered as a pixel classification problem and a novel neural architecture is employed to classify each pixel into cytoplasm, nucleus or background region. The network tuned for a set of images works well for other similar stained blood images.

Results: Comparative analysis with other standard techniques reveals that the proposed scheme outperforms its counterparts in terms of nucleus and cytoplasm extraction.

Conclusions: In this work, a neural network based lymphocyte image segmentation scheme is designed for automated leukemia detection. Desired segmentation accuracy in terms of nucleus and cytoplasm extraction is always high in automated disease recognition system and is achieved through the proposed scheme.
KEYWORD
Leukocyte segmentation, Lab color model, Neural networks, Pixel classification, Functional expansion, Tanimoto index
FullTexts / Linksout information
Listed journal information