Advances and Applications in Statistics
Volume 29, Issue 1, Pages 1 - 31
(July 2012)
|
|
APPLICATION OF GENETIC ALGORITHM FOR CLASSIFICATION OF MEDICAL IMAGES
Asanao Shimokawa and Etsuo Miyaoka
|
Abstract: Medical image processing and classification involves multiple processes that can be classified into the following general steps: preprocessing step, feature extraction step, feature selection step, and classification step. In this study, we examine the availability of medical image classification techniques that involve the processing and pattern recognition for gray images by CT scan of lung cancer patients with different grades of the disease. In the feature extraction step, features of interest were extracted from whole images as well as regions of interest by using statistical texture patterns and wavelet transformation. In the feature selection step, we used a genetic algorithm to search for a feature set that could be used for the effective classification of the images. However, in order to optimize the classification, weighting each feature helped decide the exclusion/ inclusion of features. Several genetic-algorithm-based methods for determining the optimal weight of the feature space have previously been proposed. In this article, we propose a new method for weighting each feature; the method involves the use of iterations of a genetic algorithm. The method was compared with another method. |
Keywords and phrases: |
|
Number of Downloads: 357 | Number of Views: 1204 |
|