Query by Examples for Large Image Databases

The objective of this project is to retrieve images from large image database using query by examples. Our system uses the theories of optimal projection for optimal feature selection and a hierarchical structure for low computational complexity. The system can proceed under a supervised, unsupervised, or hybrid learning mode for building the image database. In the supervised mode, a hierarchy of class labels is provided with each training image. No class labels are given under the unsupervised learning mode, and some training images are labeled in the hybrid mode. Since query by examples is closely related to recognition from image views, this project is in conjunction with the SHOSLIF-O project.

References

Dan Swets and John J. Weng, ``Using Discriminant Eigenfeatures for Image Retrieval,'' Technical Report CPS-96-16, Department of Computer Science, MSU, March 1996. Accepted by PAMI, special issue on Digital libraries: Representation and Retrieval, 1996. Click here to down load the paper (PostScript).

Dan Swets and John J. Weng, ``Hierarchical Discriminant Analysis for Image Retrieval,'' Technical Report CPS-96-17, Department of Computer Science, MSU, March 1996. Click here to down load the paper (PostScript).

Dan Swets and John J. Weng, ``HOSLIF-O: SHOSLIF for Object Recognition and Image Retrieval (Phase II),'' Technical Report CPS-95-39, Department of Computer Science, MSU, December 1995. Click here to view the paper.

D. L. Swets and J. Weng, ``Image-based recognition using learning for generalizing parameters,'' in Proc. 2nd Asian Conf. on Computer Vision, Singapore, Dec. 5-8, 1995.

D. L. Swets and J. Weng, ``Efficient content-based image retrieval using automatic feature selection,'' in Proc. IEEE Int'l Symposium on Computer Vision, Coral Gables, FL, Nov. 20-22, 1995.

Dan Swets and John J. Weng, ``Efficient image retrieval using a network with complex neurons,'' in Proc. IEEE Int'l Conf. on Neural Networks, Perth, Australia, Nov. 27 - Dec 1, 1995.

Dan Swets and John J. Weng, ``SHOSLIF-O: SHOSLIF for Object Recognition (Phase I),'' Technical Report CPS-94-64, Department of Computer Science, MSU, December 1994.

To Weng's Home Page: http://web.cps.msu.edu/~weng/