Decentralized Machine Learning Intelligence > Topics > Agent-based Pattern Recognition

Agent-based Pattern Recognition

For pattern recognition operates in real-world dynamic environments, such as remote plenary space, distributed network systems, and cloud computing mechanism, agent-based pattern recognition system accommodates the demands of a real-time dynamic environment, provides adaptable and compatible frameworks, and allows access to pattern recognition related services and applications regardless of time and location restrictions. Our researches on agent-based pattern recognition include, security service oriented multi-agent network data stream classification outlier detection, and agent based computer vision and image understanding. Figure 1 illustrates a present research on, agent based computer vision for image change detection, where an intelligent agent, modeled as an incremental learning system, is detecting change of one image to another. The purpose of this research is environmental change detection for pollution detection, examination, and prediction

tl_files/images/agent-based.pngFigure 1 Illustration of agent based computer vision for image change detection, where an intelligent agent is detecting change of one image to another.