| 摘要 |
One aspect of the invention is directed to a search engine having indexed retrieval to improve computational efficiency of searching large databases of rich objects such as images. Feature vectors are extracted from images and stored in a feature vector database. When a query is submitted to the engine, a query feature vector Q will be specified, as well as a distance threshold T, indicating the maximum distance that is of interest for this query. All images within a distance of T will be identified by the query. Range constraints are defined such that all feature vectors within a distance of T of the query feature vector, satisfy all of the range constraints. The constraint is dependent on the specific primitive being indexed. The constraint is also defined such that any feature vector which is within a distance of T of the query feature vector also satisfies a functional constraint. By reducing the number of feature vectors retrieved from the database and the number of feature vector comparisons, the query process becomes much more efficient.
|
