The Matching Algorithm using Block Segmentation for Object Recognition

The object recognition using image is one of the major concerned fields of the computer vision and still many researches are done. The feature based methods as a model based object recognition among several methods for object recognition applied to find out inherent feature point and recognize the object through feature point matching between the query images of stored in the database and the input images. But, the feature based object recognition has suffered of difficulties in real-time processing due to heavy calculations and problems that responds sensitive for change of parameter. This paper has focused on speed improvement in these problems, and proposing the methods of object recognition by dividing the images into uniform blocks and matching each blocks to identify the object for more efficient performance than the feature based object recognition. In this paper, it is confirmed that the matching time for object recognition has reduced significantly since only using minimal blocks in the proposed method.