A new fusion model for classification of the lung diseases using genetic algorithm

Automatic classification of lung diseases in computed tomography (CT) images is an important diagnostic tool for computer-aided diagnosis system. In this study, we propose a new image based feature extraction technique for classification of lung CT images. A novel fusion based method was developed by combining the Gabor filter and Walsh Hadamard transform features using median absolute deviation (MAD) technique and hence, it possesses the advantages of both models. The proposed system comprises of three stages. In the first stage, the images are preprocessed and features are extracted by novel fusion based feature extraction technique, followed by second stage, in which extracted features are selected by applying genetic algorithm which selects the top ranked features. In the final stage, classifiers namely decision tree, K nearest neighbor (KNN), Multi layer perceptron Neural Networks (MLP-NN) are employed to perform classification of the lung diseases. A total of 400 datasets for the diseases bronchitis, emphysema, pleural effusion and normal lung were used for training and testing. The classification accuracy of above 90% is accomplished by multilayer perceptron neural network classifier. The system has been tested with a number of real Computed Tomography lung images and has achieved satisfactory results in classifying the lung diseases.