English

Computationally Efficient Neural Image Compression

Image and Video Processing 2019-12-19 v1 Machine Learning Machine Learning

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.

Keywords

Cite

@article{arxiv.1912.08771,
  title  = {Computationally Efficient Neural Image Compression},
  author = {Nick Johnston and Elad Eban and Ariel Gordon and Johannes Ballé},
  journal= {arXiv preprint arXiv:1912.08771},
  year   = {2019}
}

Comments

In submission to a conference

R2 v1 2026-06-23T12:50:05.629Z