This dissertation presents several tasks that have been completed in order to achieve the above goal. First, this dissertation presents algorithms for processing color-based queries based on the colors contained within an image. They process queries of the type "Identify all images that are between PCTmin and PCTmax percent of color CQ", where PCTmin and PCTmax represent percentages and C Q represents a color in the RGB (Red, Green, Blue) model.

Third, this dissertation proposes a data structure for organizing virtual images identifiers stored in the MMDBMS in order to reduce the amount of time it takes to process the above algorithm. By using the data structure, the system will be able to identify some of the virtual images that can satisfy a given query without analyzing their sequences of editing operations. The reduction in the query processing time occurs from the reduction in the number of virtual images that have to be analyzed.

Next, this dissertation proposes algorithms for measuring the similarity between two images when one of them is stored virtually, where the similarity is based on the colors contained within an image. This allows an MMDBMS to process color-based searching queries of the type "Identify the k images that most resemble Q based on color", where k represents the desired number of images, and Q represents a query object by providing a method to measure how similar each virtual image is to the query object.

Previous research has demonstrated that instead of storing images in a Multimedia DataBase Management System (MMDBMS) using a conventional binary format, space can be saved by storing some of the images virtually, meaning that they are stored as sequences of editing operations. Since the existing techniques for searching images by color typically assume that the images are stored in conventional binary formats, new techniques and strategies for processing the queries are needed when the images are stored virtually. The goal of this dissertation is to develop techniques for performing color-based searches of virtual images and determine their strengths and weaknesses.

Finally, this dissertation constructs a prototype system to compare the above algorithms to the conventional approach for processing color-based search queries that use images stored as binary objects. The performance evaluation is based on permanent storage space used, color-based search query processing time, insertion query processing time, as well as accuracy. The comparison results show that unlike the alternative approaches, the proposed algorithms are able to perform efficiently in both searching and insertion time while still saving storage space through the use of virtual images.