Image Search Methods
– Image search is a specialized data search used to find images.
– Query terms such as keywords, image file/link, or clicking on an image can be used for searching.
– The system returns images similar to the query based on criteria like meta tags, color distribution, and region/shape attributes.
– Image meta search allows searching based on associated metadata such as keywords and text.
– Content-based image retrieval (CBIR) retrieves images based on their contents, such as textures, colors, and shapes, without relying on textual descriptions.
Data Scope
– Understanding the scope and nature of image data is crucial for designing an image search system.
– Search data can be classified into categories such as archives, domain-specific collections, enterprise collections, personal collections, and web images.
– Archives usually contain large volumes of structured or semi-structured homogeneous data.
– Domain-specific collections provide access to controlled users with specific objectives, like biomedical and satellite image databases.
– Enterprise collections are heterogeneous collections accessible within an organization’s intranet, with images stored in various locations.
Evaluations
– Evaluation workshops aim to investigate and improve the performance of image retrieval systems.
– ImageCLEF is a track of the Cross Language Evaluation Forum that evaluates systems using both textual and pure-image retrieval methods.
– Content-based Access of Image and Video Libraries is a series of IEEE workshops from 1998 to 2001.
– These evaluations contribute to advancing the field of image retrieval and enhancing system capabilities.
Related Concepts
– Automatic image annotation is a related concept in image retrieval.
– Computer vision plays a significant role in image retrieval systems.
– Concept-based image indexing is another approach to organizing and retrieving images.
– Digital asset management involves the storage and organization of digital images.
– Digital image editing is often performed on retrieved images for various purposes.
References and Resources
– B E Prasad, A Gupta, H-M Toong, and S.E. Madnick developed the first microcomputer-based image database retrieval system at MIT in the 1990s.
– A 2008 survey article by Datta, Joshi, Li, and Wang documented the progress in image retrieval systems.
– Camargo, Caicedo, and Gonzalez proposed a kernel-based framework for image collection exploration in 2013.
– The references provide further reading and resources for understanding image retrieval.
– External links like Image-Net.org and VGG Image Search Engine offer additional tools and information related to image retrieval.
An image retrieval system is a computer system used for browsing, searching and retrieving images from a large database of digital images. Most traditional and common methods of image retrieval utilize some method of adding metadata such as captioning, keywords, title or descriptions to the images so that retrieval can be performed over the annotation words. Manual image annotation is time-consuming, laborious and expensive; to address this, there has been a large amount of research done on automatic image annotation. Additionally, the increase in social web applications and the semantic web have inspired the development of several web-based image annotation tools.
The first microcomputer-based image database retrieval system was developed at MIT, in the 1990s, by Banireddy Prasaad, Amar Gupta, Hoo-min Toong, and Stuart Madnick.
A 2008 survey article documented progresses after 2007.
All image retrieval systems as of 2021 were designed for 2D images, not 3D ones.
