A Multi-Agents Architecture to Learn Vision Operators and their Parameters
This addresses the problem for users in vision systems who struggle with operator and parameter selection, though it appears incremental as it builds on existing multi-agent and reinforcement learning approaches.
The paper tackles the challenge of selecting and tuning vision operators and their parameters by proposing a multi-agent architecture that learns optimal combinations for a class of images, resulting in improved computation time and result quality through reinforcement learning.
In a vision system, every task needs that the operators to apply should be « well chosen » and their parameters should be also « well adjusted ». The diversity of operators and the multitude of their parameters constitute a big challenge for users. As it is very difficult to make the « right » choice, lack of a specific rule, many disadvantages appear and affect the computation time and especially the quality of results. In this paper we present a multi-agent architecture to learn the best operators to apply and their best parameters for a class of images. Our architecture consists of three types of agents: User Agent, Operator Agent and Parameter Agent. The User Agent determines the phases of treatment, a library of operators and the possible values of their parameters. The Operator Agent constructs all possible combinations of operators and the Parameter Agent, the core of the architecture, adjusts the parameters of each combination by treating a large number of images. Through the reinforcement learning mechanism, our architecture does not consider only the system opportunities but also the user preferences.