Welcome to the FAST image retrieval technology


Developed at
Multimedia Research Laboratory
Binghamton University
State University of New York
Copyright © 2004



This demonstration is about the Fast And Semantics-Tailored (FAST) image retrieval methodology we have developed. Specifically, the contributions of FAST methodology to the CBIR literature include: (1) development of a new indexing method based on fuzzy logic to incorporate color, texture, and shape information into a region based approach to improving the retrieval effectiveness and robustness (2) development of a new hierarchical indexing structure and the corresponding Hierarchical, Elimination-based A* Retrieval algorithm (HEAR) to significantly improve the retrieval efficiency without sacrificing the retrieval effectiveness; it is shown that HEAR is guaranteed to deliver a logarithm search in the average case (3) employment of user relevance feedbacks to tailor the semantic retrieval to each user's individualized query preference through the novel Indexing Tree Pruning (ITP) and Adaptive Region Weight Updating (ARWU) algorithms. Theoretical analysis and experimental evaluations show that FAST methodology holds a great promise in delivering fast and semantics-tailored image retrieval in CBIR. This work is published at: 

Ruofei Zhang and Zhongfei (Mark) Zhang, FAST: Towards More Effective and Efficient Image Retrieval, ACM Multimedia Systems Journal, special issue on Multimedia Information Retrieval, Springer, Vol. 10, Number 6, October, 2005, pp 529-543 

This demonstration offers two options for users to query the image database. The first option allows a user to query the database using his/her own images in their local file systems. The second option allows a user to randomly browse the image database and pick up one image to query the database. In both scenarios, the user is offered the relevance feedback interactions for each round of retrieval. The user may casts relevant (by clicking +), irrelevant (by clicking -), or neutral (which is default with no action). The user does not need to label all the images in each retrieval.

A further research on the relevance feedback addressing the small sample issue and the asymmetric sample issue simultaneously is done with the development of the BALAS methodology. The detailed work is published at: 

Ruofei Zhang and Zhongfei (Mark) Zhang, Empirical Bayesian Learning in the Relevance Feedback of Image Retrieval, Image and Vision Computing, Elsevier Science, Volume 24, Issue 3, March, 2006, pp 211-223





The usage of the buttons on this page

Option 1
Upload query image file

Click the Browse button to choose an image file in the local directory
Click the Preview button to view the file you choose (optional)
Upload and query by clicking the Submit button
Option 2
Choose radom file
Click Random button to get image randomly from database
Random

Developed by Qixiong Zhen (2005)



This material is based upon the work supported by Binghamton University, New York State, and Microsoft Research through a faculty visiting researcher fellowship and a graduate student internship.

Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the funding agencies.

Go back to the Multimedia Computing Research Lab homepage