DEEP IMAGE COMPRESSION SYSTEM FOR THE BETTER RATE-DISTORTION PERFORMANCE

Abstract

This paper analyses the deep image compression system that contains encoder,
quantizer, entropy model and decoder optimized by joint rate-distortion framework.
Current model implements a channel-level variable quantization network to dynamically
allocate and withdraw the bitrates from significant and negligible channels. Its main
specific is usage of the variable quantization controller that consists of such components:
channel importance module that dynamically learns the importance of channels during the
training, and splitting-merging module, which allocates different bitrates of the channels.
Quantizer implements the Gaussian mixture model manner. The paper is a continuation of
several similar research done before that first provided an idea and architecture of the
model. The main goals of the proposed work are to do deeper analysis of the system, verify
and improve the model effectiveness. The experiments validate that hyper-parameter
tuning approach proposed here successfully improves the rate-distortion performance of
image compression in terms of various quality metrics such as PSNR, MS-SSIM and BPP.
Key words: convolutional neural networks, Deep learning, Deep Image Compression,
Channel-Level Variable Quantization, Deep Neural Networks.
Fig.: 3, Tabl.: 3, Bibl.: 11.