Video indexing and image retrieval system

DETAILED DESCRIPTION The present invention relates generally to video signal processing.
More particularly, the invention relates to a video indexing and image retrieval system.
Over the last few years, there has been a dramatic increase in available bandwidth over high-speed networks.
At the same time, computer manufacturers improved the storage capacities of hard drives on personal computers, and improved the speed of the system bus and the motherboards that access the hard drives.
The quality and efficiency of data compression algorithms has likewise improved information transmission efficiency and access rates—particularly with respect to video data.
One of the most important tasks of a database manager is to provide easy and intuitive access to data.
This task can be particularly difficult when a user would like to search for images or other visual data such as video segments.
Browsing and searching for data is one useful way that the Internet allows users to access related Internet pages rapidly and intuitively through text-based searches.
To allow the searching of visual data, an image retrieval system must be able to emphasis the similarity of a query image with images or frames of video stored in a database.
There are several ways that a user may provide a query image.
For example, users may have a rough idea of an image that they are looking for.
The user may develop a simple sketch of an image by hand and a scanner can be used to upload the sketch or drawing software can be used.
A photo of the image or a similar image can be used to find other similar images in the database.
An image search engine must be able to generate a measurement of the similarity between the query image and database images so that the user is presented with a list of the most relevant database images to the least relevant images.
The image search engine associated with the image retrieval system must be able to look for similarities between significant features of the query sketch or image and the database images while ignoring minor detail variations