Graphical User Interface for Image Retrieval Applications

Technology

This invention is related to the fields of image processing and information retrieval. There exists three broad categories of image retrieval systems: those that are content-based, those that are content-free (employ collaborative filtering), and those that rely on manual text annotation. Each of these approaches has unique advantages and distinct limitations. As many of the limitations are mutually exclusive, a hybrid approach that is able to combine these retrieval methods holds promise.

Florida Atlantic University inventors have developed a novel web-based interface known as the Perceptually-Relevant Image Search Machine (PRISM). The unique interface design allows the simultaneous execution of the three distinct retrieval methods through a single query. Content-based, collaborative, and text similarity results are all generated from one query within a single interface. Furthermore, the system learns from each query, improving its performance over time. The invention provides a new way to organize and retrieve physically or semantically relevant images.

Applications

• Medical imaging (image retrieval, image analysis)
• Surveillance and video surveillance (law enforcement, security)
• Publishing (photography, print media, advertising)
• Real estate
• Consumers (digital photo organization

Advantages

• Enables a user to group, scale, and annotate images. Images can be spatially organized by placing related images together or in close proximity. It is also possible to scale images so that the relative size of an image reflects its relevance to the user.
• A tabbed interface provides the user with the capability to group images in multiple, distinct collections.
• Multiple users can simultaneously and independently organize images, while taking advantage of the others’ work.
• Information gathered from multiple users and across sessions can be pooled for improved results.