A Modified and Memory Saving Approach to B+ Tree Index for Search of an Image Database based on Chain Codes

Space savings is a demand of these days. Almost all applications are trying to represent their data in the least possible amount of space. It is the goal of all algorithms to consume as less memory as possible in the computer. Many techniques have been developed to compress large bulk of data so that large amount of information can be accommodated in lesser amount of space. The proposed approach called Coded B+ tree is a modified approach to the conventional B+ Tree index and can be applied only to chain codes of images and to a search of an image database based on chain codes of images as its search key value. The proposed method brings about considerable space savings of more than a half at the cost of 4 to 5 more comparisons. It is a useful technique of indexing for systems that require saving memory spaces. General Terms Image data base indexing, Digital Image Processing, Content Based Image Retrieval.