The approach requires the archive images to be associated with an index pattern. The encoding process is determined by holographic learning. This learning algorithm generates an Holograph. The retreival process uses only this holograph.
During retrieval, a sample image is first accepted from the searcher. The searcher can also specify one or more particular object(s) in this sample as a focul index. This information (sample image + focus mask) is then used to generate a special search template. This template is then "resonated" with the Holograph by holograph decoding algorithm. This is computationally an single step convolution operation. The output is an index pattern. This index pattern corresponds to the closest match and directly identifies the matching image. Below, an sample database is given and some query example is shown.