期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2014
卷号：67
期号：1
出版社：Journal of Theoretical and Applied
摘要：The object detection is a very important technique in computer vision. This detection is mainly used in many applications like military, satellite image mining, medical and etc. This paper proposes the object detection based on data clustering methods in region based segmentation and shape feature. In existing, most of researchers are using the k-means and fuzzy k-means for clustering and it uses the SVM and Adaboost classifiers for object classification. Here each cluster needs the own centric and distance calculation for clustering. The main disadvantage of this technique is distance calculation between the pixels. This distance calculation technique does not produce the efficient result in clustering. In classification, SVM classifier needs more parameters for increasing the efficiency and adaboost is more noise sensitivity. To avoid these drawbacks, the region based segmentation using non Euclidean distance measure for clustering and combined Fisher SVM with modified adaboost algorithm is used for object classification. This result shows the region based segmentation and object classification of an image. Finally, the performance analysis graph shows the increased efficiency of proposed algorithm.
