Publications

Abstract : Green Bio Technology is the most focused research area in image processing for improving the process of agriculture. Herewith image classification is used to solve the problems identified in green bio technology. Classification is a process of grouping the similar data based on its features. Images are having the different types of features, i.e. Texture, shape and color. Normally, image classification is performed using Feature extraction, Selection and similarity measurements. In classification, most of the researchers are using the Gray level Co Occurrence Matrix (GLCM) for texture feature extraction. The main drawback of existing algorithm of GLCM is global features generation. To overcome the above mentioned drawback, this paper proposes Density clustering based GLCM (DGLCM) for extracting the local features of an in image. When compared to global, the local features are having more information of an image. The similarities of features are calculated using Euclidean distance measure. Finally, the features are classified using Fisher kernel based Support Vector Machine (FSVM).The performance of this algorithm is calculated using precision and recall. The result is evaluated using the images collected from the agricultural processes.