The JPEG still picture compression standard
Communications of the ACM - Special issue on digital multimedia systems
Communications of the ACM
| Hi-index | 0.00 |
Many current multimedia database management systems perform content-based retrieval of images by extracting a set of features from each data object inserted into the underlying database. This allows users to search the database for images that are similar to an input query image where the similarity is determined by comparing the previously extracted features. Thus, the ability of the system to compare images depends upon the features it can extract from them. This also means that the system may fail to correctly match one of its stored images to the query image if their corresponding features do not sufficiently match. One approach to minimize the occurrences of these situations is to add modified versions of the stored image to the database and attempt to match the query image to any one of them. This paper discusses the advantages and applicability of this approach.