Microdosing: Knowledge Distillation for GAN based Compression
This work addresses the practical limitations of GAN-based compression for real-world applications, offering an incremental improvement in efficiency.
The paper tackles the problem of large model size and high computational cost in GAN-based image and video compression by applying knowledge distillation to create smaller decoders, achieving a 20x reduction in model size and a 50% reduction in decoding time.
Recently, significant progress has been made in learned image and video compression. In particular the usage of Generative Adversarial Networks has lead to impressive results in the low bit rate regime. However, the model size remains an important issue in current state-of-the-art proposals and existing solutions require significant computation effort on the decoding side. This limits their usage in realistic scenarios and the extension to video compression. In this paper, we demonstrate how to leverage knowledge distillation to obtain equally capable image decoders at a fraction of the original number of parameters. We investigate several aspects of our solution including sequence specialization with side information for image coding. Finally, we also show how to transfer the obtained benefits into the setting of video compression. Overall, this allows us to reduce the model size by a factor of 20 and to achieve 50% reduction in decoding time.