A GAN-based Tunable Image Compression System
This addresses a bottleneck in efficient image compression systems, offering a tunable scheme for specific compression ratios without retraining, though it is incremental as it builds on existing GAN-based methods.
The paper tackles severe distortion in non-important regions of content-weighted image compression at low bits per pixel by using a GAN for reconstruction and multiscale pyramid decomposition, achieving over 10.3% improvement in MS-SSIM at 0.05 bpp on the Kodak dataset.
The method of importance map has been widely adopted in DNN-based lossy image compression to achieve bit allocation according to the importance of image contents. However, insufficient allocation of bits in non-important regions often leads to severe distortion at low bpp (bits per pixel), which hampers the development of efficient content-weighted image compression systems. This paper rethinks content-based compression by using Generative Adversarial Network (GAN) to reconstruct the non-important regions. Moreover, multiscale pyramid decomposition is applied to both the encoder and the discriminator to achieve global compression of high-resolution images. A tunable compression scheme is also proposed in this paper to compress an image to any specific compression ratio without retraining the model. The experimental results show that our proposed method improves MS-SSIM by more than 10.3% compared to the recently reported GAN-based method to achieve the same low bpp (0.05) on the Kodak dataset.