Transformations in Learned Image Compression from a Modulation Perspective
This work addresses the challenge of improving compression efficiency in learned image compression for applications like storage and transmission, offering an incremental advance by applying communication theory insights.
The paper tackles the problem of designing transformation methods in learned image compression by interpreting it as a communication system, proposing a unified transform method based on signal modulation that reduces existing methods to linear modulation and extends to nonlinear variants. The result includes a 3.52% BD-rate reduction on the Kodak dataset without increased complexity.
In this paper, a unified transformation method in learned image compression(LIC) is proposed from the perspective of modulation. Firstly, the quantization in LIC is considered as a generalized channel with additive uniform noise. Moreover, the LIC is interpreted as a particular communication system according to the consistency in structures and optimization objectives. Thus, the technology of communication systems can be applied to guide the design of modules in LIC. Furthermore, a unified transform method based on signal modulation (TSM) is defined. In the view of TSM, the existing transformation methods are mathematically reduced to a linear modulation. A series of transformation methods, e.g. TPM and TJM, are obtained by extending to nonlinear modulation. The experimental results on various datasets and backbone architectures verify that the effectiveness and robustness of the proposed method. More importantly, it further confirms the feasibility of guiding LIC design from a communication perspective. For example, when backbone architecture is hyperprior combining context model, our method achieves 3.52$\%$ BD-rate reduction over GDN on Kodak dataset without increasing complexity.