Processing Energy Modeling for Neural Network Based Image Compression

TL;DR

This paper analyzes the energy consumption of neural network-based image compression on GPUs, showing it can be estimated from image size and is mainly driven by operations per pixel, especially in early network layers.

Contribution

It provides an energy modeling approach for neural image compression, highlighting the main energy drivers and estimating consumption with high accuracy.

Findings

Energy consumption correlates strongly with image size.

Operations per pixel are the main energy driver.

Early network layers consume most energy.

Abstract

Nowadays, the compression performance of neural-networkbased image compression algorithms outperforms state-of-the-art compression approaches such as JPEG or HEIC-based image compression. Unfortunately, most neural-network based compression methods are executed on GPUs and consume a high amount of energy during execution. Therefore, this paper performs an in-depth analysis on the energy consumption of state-of-the-art neural-network based compression methods on a GPU and show that the energy consumption of compression networks can be estimated using the image size with mean estimation errors of less than 7%. Finally, using a correlation analysis, we find that the number of operations per pixel is the main driving force for energy consumption and deduce that the network layers up to the second downsampling step are consuming most energy.