Selective Perception and Applications

We are developing a novel scheme for efficient compression and description of the semantic content of video and images.
The compression is based on content-correlates in a sense that we find the regions of interest (ROIs) and then compress the "sub-images" within the ROIs.
The compression scheme uses wavelet transform to decompose an image into the frequency subbands, then searches for the regions of interest within each subband. You can retrieve a paper describing in detail the region finder. The resulting regions are ranked by "importance" and transmitted in this order. We are working on extending the scheme to achieve content-based compression using:

• recognition of the content of regions of interest
• matching the invariant sub-graphs of ROI relationships

Publications

1. Ali Shokoufandeh, Ivan Marsic, Sven J. Dickinson, View-Based Object Recognition Using Saliency Maps,   Image and Vision Computing, Vol.17, Issues 5-6, pp.445-460, April 1999.
   (doi:   10.1016/S0262-8856(98)00124-3)
2. Ali Shokoufandeh, Ivan Marsic and Sven J. Dickinson, View-Based Object Matching,   In Proceedings of the IEEE International Conference on Computer Vision (ICCV'98), pages 588-595, Bombay, India, January 1998.
   (doi:   10.1109/ICCV.1998.710777)
3. Ali Shokoufandeh, Ivan Marsic, Sven J. Dickinson, View-Based Object Recognition Using Saliency Maps, Technical Report DCS-TR-339, Department of Computer Science, Rutgers University.
4. Thomas Slowe and Ivan Marsic, Saliency-Based Visual Representation for Compression
   in Proc. IEEE Int'l Conf. Image Processing (ICIP'97), Santa Barbara, CA, Vol.2, pages 554-557, October 1997.
   (doi:   10.1109/ICIP.1997.638831)
5. Ali Shokoufandeh, Ivan Marsic and Sven J. Dickinson, Saliency Regions as a Basis for Object Recognition, In Proceedings of the 3rd International Workshop on Visual Form (IWVF3), pages 539-548, Capri, Italy, May 1997.
6. Ivan Marsic, Evidential Reasoning in Visual Recognition, In Intelligent Engineering Systems Through Artificial Neural Networks 4, C.H. Dagli, B.R. Fernandez, J. Ghosh, and R.T.S. Kumara (eds.), ASME Press, New York, pp.511-516, 1994.
7. Ivan Marsic, Data-Driven Shifts of Attention in Wavelet Scale Space, CAIP-TR-166, CAIP Center, Rutgers University, September 1993. (IEEE Trans. PAMI, submitted)
8. Evangelia Micheli-Tzanakou and Ivan Marsic, Visual Shape Analysis by Wavelets, In Proceedings of the RNNS/IEEE Symposium on Neuroinformatics and Neurocomputing, pages 289-298, Rostov-on-Don, Russia, October 1992.
9. Ivan Marsic and Evangelia Micheli-Tzanakou, A Framework for Object Representation and Recognition, In Proceedings of the IEEE International Joint Conference on Neural Networks (IJCNN '92), Vol.3, pages 272-277, Baltimore, MD, June 1992.
   (doi:   10.1109/IJCNN.1992.227162)
10. Evangelia Micheli-Tzanakou and Ivan Marsic, Object Detection and Recognition in a Multiresolution Representation, In Proceedings of the IEEE International Joint Conference on Neural Networks (IJCNN '91), Vol. 3, pages 2564-2569, Singapore, November 1991.
11. Ivan Marsic and Evangelia Micheli-Tzanakou, Distributed Optimization with the ALOPEX Algorithms, In Proceedings of the Twelfth Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBS), vol. 12, pp. 1415-1416, 1990.

Related Work

Back to CAIP Multimedia Page