Computationally Efficient Neural Image Compression

Abstract

Image compression using neural networks have reached or exceeded non-neural methods (such as JPEG, WebP, BPG). While these networks are state of the art in ratedistortion performance, computational feasibility of these models remains a challenge. We apply automatic network optimization techniques to reduce the computational complexity of a popular architecture used in neural image compression, analyze the decoder complexity in execution runtime and explore the trade-offs between two distortion metrics, rate-distortion performance and run-time performance to design and research more computationally efficient neural image compression. We find that our method decreases the decoder run-time requirements by over 50% for a stateof-the-art neural architecture.

In submission to a conference

References in corpus (4)

• End-to-end Optimized Image Compression
• Joint Autoregressive and Hierarchical Priors for Learned Image Compression
• Learning the Number of Neurons in Deep Networks
• Auto-Sizing Neural Networks: With Applications to n-gram Language Models

Cited by in corpus (13)

• Improving Inference for Neural Image Compression
• Hierarchical Quantized Autoencoders
• Modeling Lost Information in Lossy Image Compression
• Checkerboard Context Model for Efficient Learned Image Compression
• End-to-end optimized image compression with competition of prior distributions
• CBANet: Towards Complexity and Bitrate Adaptive Deep Image Compression using a Single Network
• End-to-End Learned Image Compression with Quantized Weights and Activations
• Slimmable Compressive Autoencoders for Practical Neural Image Compression
• Variable-Rate Deep Image Compression through Spatially-Adaptive Feature Transform
• Learned Block-based Hybrid Image Compression
• Improving Lossless Compression Rates via Monte Carlo Bits-Back Coding
• Exploring Autoencoder-based Error-bounded Compression for Scientific Data
• DeepCompress: Efficient Point Cloud Geometry Compression