Variable Rate Image Compression with Recurrent Neural Networks

Abstract

A large fraction of Internet traffic is now driven by requests from mobile devices with relatively small screens and often stringent bandwidth requirements. Due to these factors, it has become the norm for modern graphics-heavy websites to transmit low-resolution, low-bytecount image previews (thumbnails) as part of the initial page load process to improve apparent page responsiveness. Increasing thumbnail compression beyond the capabilities of existing codecs is therefore a current research focus, as any byte savings will significantly enhance the experience of mobile device users. Toward this end, we propose a general framework for variable-rate image compression and a novel architecture based on convolutional and deconvolutional LSTM recurrent networks. Our models address the main issues that have prevented autoencoder neural networks from competing with existing image compression algorithms: (1) our networks only need to be trained once (not per-image), regardless of input image dimensions and the desired compression rate; (2) our networks are progressive, meaning that the more bits are sent, the more accurate the image reconstruction; and (3) the proposed architecture is at least as efficient as a standard purpose-trained autoencoder for a given number of bits. On a large-scale benchmark of 32$\times$32 thumbnails, our LSTM-based approaches provide better visual quality than (headerless) JPEG, JPEG2000 and WebP, with a storage size that is reduced by 10% or more.

Under review as a conference paper at ICLR 2016

References in corpus (8)

• Sequence to Sequence Learning with Neural Networks
• Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting
• Recurrent Neural Network Regularization
• BinaryConnect: Training Deep Neural Networks with binary weights during propagations
• Deep Generative Image Models using a Laplacian Pyramid of Adversarial Networks
• DRAW: A Recurrent Neural Network For Image Generation
• Fully Convolutional Networks for Semantic Segmentation
• Techniques for Learning Binary Stochastic Feedforward Neural Networks

Cited by in corpus (51)

• Wireless Image Transmission Using Deep Source Channel Coding With Attention Modules
• Soft-to-Hard Vector Quantization for End-to-End Learning Compressible Representations
• Lossy Image Compression with Compressive Autoencoders
• High-Fidelity Generative Image Compression
• Deep Learning-Based Video Coding: A Review and A Case Study
• Deep Contextual Video Compression
• Real-Time Adaptive Image Compression
• Towards Image Understanding from Deep Compression without Decoding
• CLVSA: A Convolutional LSTM Based Variational Sequence-to-Sequence Model with Attention for Predicting Trends of Financial Markets
• Learned Image Compression with Discretized Gaussian Mixture Likelihoods and Attention Modules
• Adversarial examples for generative models
• BlockCNN: A Deep Network for Artifact Removal and Image Compression
• Generalized Octave Convolutions for Learned Multi-Frequency Image Compression
• Learning Convolutional Networks for Content-weighted Image Compression
• FenceBox: A Platform for Defeating Adversarial Examples with Data Augmentation Techniques
• Deep Generative Models for Distribution-Preserving Lossy Compression
• Image compression optimized for 3D reconstruction by utilizing deep neural networks
• Distributed Learning and Inference with Compressed Images
• End-to-End Rate-Distortion Optimization for Bi-Directional Learned Video Compression
• BINet: a binary inpainting network for deep patch-based image compression
• A Cross Channel Context Model for Latents in Deep Image Compression
• Learning to Compress Videos without Computing Motion
• Validation and parameterization of a novel physics-constrained neural dynamics model applied to turbulent fluid flow
• DSSLIC: Deep Semantic Segmentation-based Layered Image Compression
• Deep Implicit Volume Compression
• Learning Content-Weighted Deep Image Compression
• Learning a Single Tucker Decomposition Network for Lossy Image Compression with Multiple Bits-Per-Pixel Rates
• CAE-ADMM: Implicit Bitrate Optimization via ADMM-based Pruning in Compressive Autoencoders
• Slimmable Compressive Autoencoders for Practical Neural Image Compression
• Perceptual Quality Study on Deep Learning based Image Compression
• Deep Convolutional AutoEncoder-based Lossy Image Compression
• CompressNet: Generative Compression at Extremely Low Bitrates
• Learned Scalable Image Compression with Bidirectional Context Disentanglement Network
• Neural Multi-scale Image Compression
• Attention Based Image Compression Post-Processing Convolutional Neural Network
• Learning to Structure an Image with Few Colors
• Learned Variable-Rate Image Compression with Residual Divisive Normalization
• Learned Lossless Image Compression with a HyperPrior and Discretized Gaussian Mixture Likelihoods
• DeepSIC: Deep Semantic Image Compression
• End-to-End Image Compression with Probabilistic Decoding
• Substitutional Neural Image Compression
• Attention-guided Image Compression by Deep Reconstruction of Compressive Sensed Saliency Skeleton
• Deep Learning-based Image Compression with Trellis Coded Quantization
• Learn to Compress CSI and Allocate Resources in Vehicular Networks
• Learning Image and Video Compression through Spatial-Temporal Energy Compaction
• Object Detection-Based Variable Quantization Processing
• Towards Modality Transferable Visual Information Representation with Optimal Model Compression
• Low Bitrate Image Compression with Discretized Gaussian Mixture Likelihoods
• Learned Multi-Resolution Variable-Rate Image Compression with Octave-based Residual Blocks
• Learning to Localize Through Compressed Binary Maps
• Deep motion estimation for parallel inter-frame prediction in video compression