Optimization design of biorthogonal filter banks for image compression

Abstract

Abstract In this paper, we present a new approach for designing filter banks for image compression. This approach has two major components: optimization and generalization. In the optimization phase, we formulate the design problem as a nonlinear optimization problem whose objective consists of both the performance metrics of the image coder, such as the peak signal-to-noise ratio (PSNR), and those of individual filters. Filter banks are optimized in the optimization phase based on a set of training images. In the generalization phase, the filter bank that can be generalized to other images is selected from the candidates obtained in the optimization phase to be the final result. The filter bank selected should perform well not only on the training examples used in the design process but also on test cases not seen. In contrast to existing methods that design filter banks independently from the other operations in an image compression algorithm, our approach allows us to find filter banks that work best in a specific image compression algorithm for a certain class of images. In system prototype development, we adopt the agent-based approach to achieve better modularity, portability, and scalability. Agents in the multi-agent system are specialized in performing problem formulation, image compression, optimization, and generalization. In the experiments, we show that on a set of benchmark images our approach has found filter banks that perform better than the existing filter banks in different image compression algorithms and at different compression ratios.