Article Summary #6: Image Retrieval, a LS 566 Post

From the Article

Abebe Rorissa discusses in Image Retrieval: Benchmarking Visual Information Indexing and Retrieval Systems how, even in this electronic age, a picture is worth a thousand words. People shares images and videos online on many websites. Because of this popularity he poses the question: What is the current status of images and video indexing and retrieval? He covers three main issues: firstly, efforts have been made to automate the process but it is highly limited; secondly, there are a lot of images out there but little organization with it; thirdly, there are challenges in analyzing the images that come from various domains. Rorissa goes over two approaches to indexing and retrieval: concept-based and content-based. Concept-based relies on people to manually index while content-based is automated using color, texture or shape to organize. He stresses that a combination of the two should be adopted as the best approach in indexing and retrieval. He also touches on how for videos query-by-content is the preferred retrieval method. A lot of strives have been made but improvements have to be continuously made to make indexing and retrieval better.

This article was written ten years ago to the month. As I've mentioned before that is in dog years for anything electronic. But the points that Rorissa makes are still valid today. Improvements have been made since this publication and improvements are still continuously being made. I am guilty of what he mentioned, taking lots of pictures, dumping them on my computer, then taking more pictures without organizing them. My computer organizes my pictures by the date they were taken so it's not a complete mess. It will be interesting to see how this will develop in the future and maybe someday there will be a standardized method like Dewey for books, but for images and videos.