Academic Summary

2002-01-15

(A detailed version of 2000)

My work in Image Processing can be divided into two branches. The first is Image Database. I have a lot of work in Medical Image Database, PACS and Drawing Image Processing. I proposed the Drawing Archiving and Communication System (DACS). The second branch is Pattern Recognition. I was the main research assistant in several international or national projects about associative memory, optical-electronic system and content-based retrieval. I proposed the frame algorithm of a video retrieval system based on Chinese caption and enhanced the real-time performance and robustness of the Optical-electronic Shape Recognition System.

In the field of Artificial Intelligence, with the background of Image Processing and Pattern Recognition, I think we should study the artificial intelligent system as self-organized system; a complex intelligence can be obtained by the social interaction among some simple-intelligentized individuals. I proposed some basic principles and methods in the design of intelligent systems. My theory mainly including two parts: 1. Information Theory and the definition and mathematic model of Infosys, 2.inter-relationship of Infosys or ˇ± Ecology in Computational Intelligenceˇ±.

I practice my principles and methods in Neural Network and Agent System. In the field of agent, I proposed the Fieldbus Based Agent Distributed Control System and its basic control principles in spring 2000. Furthermore, I proposed Behavior-Evolution-based Genetic Algorithm (BEGA) in 2000 and applied it in expert system and network management. In the field of Neural Network, my early work was concentrated on the analysis of associative memory. I proposed a new explain based on ˇ°energy well ˇ± for the analysis of tri-order Hopfield Network. Then I turned my interest on Synergetic Neural Network (SNN). I have done some work in the theory and application of SNN. Most important, I proposed that the competition sub-network and matching sub-network of SNN can be designed independently and order parameters can be obtained by some new methods besides traditional pseudo-inversion method. Recently I proposed the concept of Hierarchy Neural Network based on the hierarchy structure in brain. I apply this method first in SNN.

In the past year, my work focused on

  1. In-depth research on Synergetic Neural Network, including construction of the matching sub-network, entropy of SNN, Quick Haken Network and potential function analysis
  2. Using Behavior-Evolution-based Genetic Algorithm (BEGA) in information filtering of newsgroup system
  3. Using neural network and statistical methods to predict energy consumption in a power system.


[Return to Jie Bao's Homepage]