• find images with short, thin, white objects in the bottom right-hand corner;
• find images containing red ellipses arranged in a square;
• find images containing yellow regions in the center; and
• find more images that look like this one.
• scene recognition,
• object recognition, and
• knowledge discovery.
• identifying classes of objects,
• depicting image regions that might include examples of the objects, and
• giving an instance of mechanisms to validate the object presence.
• Customizing user queries. Each target application of indexing and retrieval has its own range of special needs and constraints. Systems that fail to address this requirement are unlikely to perform well enough to convince users of the technology's usefulness.
• Identifying best practices in fields that could potentially benefit from content-based multimedia indexing, including management of image collections and drawing archives, electronic publishing, and multimedia content creation.
• Addressing privacy issues and the implications on civil liberties resulting from using images, audio, and video in various applications.
Chabane Djeraba is an associate professor at the Polytechnic School of Nantes University. His research interests include design and analysis of visual information systems, content-based image and audio indexing and retrieval, and multimedia synchronization in an object-oriented database. He has a BS in computer engineering from the National Institute of Computer Science, Algeria; an MSc in computer science from Pierre Mendes France University, France; and a PhD in computer science from Claude Bernard University, France. He is an IEEE member. He is also on the Journal of Multimedia Tools and Applications editorial board.