1 Project: Efficient Management of Image and Video Data


Faculty  B.S. Manjunath ,S. K. Mitra ,Y. F. Wang

Student Wei Ma, S. K. Strobel, Y. Deng

Alexandria Digital Library Project


Netra (instructions)

Netra-II (instructions)


Search using color and texture 

Large collections of images and video pose challenging problems in storage and access to information. Our focus has been on developing multiresolution representation for storage to facilitate browsing of images, and tools for content based search of images and video. Towards this, we have developed lossless image coding algorithms for multispectral/color images using wavelet transforms that outperforms existing state-of-the art compression schemes.

Texture Feature Representation based on Gabor Wavelet Decomposition 

For content analysis, a robust texture feature set has been developed based on a Gabor wavelet decomposition, and this algorithm has been implemented in UCSB Alexandria Digital Library to provide texture-based pattern retrievals in aerial photo database. To learn more about these texture features, click on the image icon.

Learning similarity measures  

Learning similarity helps in better retrievals.

Content-based image browsing and retrieval using texture features 

A new method for indexing large image databases is proposed. This method incorporates neural network learning algorithms and pattern recognition techniques to construct an image texture dictionary. Image retrieval is then formulated as a process of dictionary search to compute the best matching codeword, which in turn indexes into the database items. This pattern dictionary integrates both learning similarity measures and efficient image indexing into one single framework. Thus, it is not only able to facilitate the search process, but also provide visually more relative retrieval results.

A texture thesaurus for browsing large aerial photographs

A texture-based image retrieval system for browsing large-scale aerial photos is presented. The salient component of this system include texture feature extraction, image segmentation and grouping, learning similarity measure, and a texture thesaurus model for fast searching and indexing. 

Related work

Image Browsing in the Alexandria Digital Library (ADL) Project, D-Lib Magazine, August 1995

Home | People | Research | Publications | Courses | Seminars | Links