Lossy Image Compression with Compressive Autoencoders
Lucas Theis, Wenzhe Shi, Andrew Cunningham、Ferenc Huszár Twitter London, UK
Abstract
We propose a new approach to the problem of optimizing autoencoders for lossy image compression. New media formats, changing hardware technology, as well as diverse requirements and content types create a need for compression algorithms which are more flexible than existing codecs. Autoencoders have the potential to address this need, but are difficult to optimize directly due to the inherent non-differentiabilty of the compression loss. We here show that minimal changes to the loss are sufficient to train deep autoencoders competitive with JPEG 2000 and outperforming recently proposed approaches based on RNNs. Our network is furthermore computationally efficient thanks to a sub-pixel architecture, which makes it suitable for high-resolution images. This is in contrast to previous work on autoencoders for compression using coarser approximations, shallower architectures, computationally expensive methods, or focusing on small images.
中文速览
有损图像压缩（lossy image compression）中，量化（quantization）步骤不可微分，导致自编码器（autoencoder）难以用梯度方法直接训练——这正是本文要解决的核心难题。作者提出了一种极简却有效的解决方案：在反向传播时用恒等函数的梯度替换量化函数的梯度，同时用可微分的高斯尺度混合密度来近似估计编码比特率，再配合基于子像素卷积（sub-pixel convolution）的高效深层网络架构，从而在不改变前向传播逻辑的前提下实现端到端优化。在 Kodak 等标准测试集上，该方法在感知质量评估指标上达到了与 JPEG 2000 相当乃至更优的压缩效果，并且明显优于同期基于循环神经网络（RNN）的压缩方案。这项工作表明，神经网络压缩编解码器完全可以在计算高效的前提下追上传统标准编解码器的性能，为面向特定内容、特定指标乃至新型媒体格式的灵活压缩框架奠定了基础。
原文 arXiv:1703.00395;中英对照 + 大白话阅读 https://aha.fim.ai/paper/1703.00395v1