Image retrieval using the longest approximate common subsequences

Accurate and fast retrieval as well as efficient storage are two principal needs in an image database. Image search methods are the crucial factors in fulfilling these requirements. The longest common subsequence search method is mostly used in string-based image databases. However, a longest common subsequence does not always reveal the degree of difference between two strings. This research proposes a new image matching method, i.e., finding a longest approximate common subsequence of two strings. The experimental results show any image in the system can be precisely retrieved as long as each image has a unique image string, where accuracy of image retrieval depends on the level of query detail.