Attention-Based Generative Neural Image Compression on Solar Dynamics Observatory

Abstract

NASA's Solar Dynamics Observatory (SDO) mission gathers 1.4 terabytes of data each day from its geosynchronous orbit in space. SDO data includes images of the Sun captured at different wavelengths, with the primary scientific goal of understanding the dynamic processes governing the Sun. Recently, endto-end optimized artificial neural networks (ANN) have shown great potential in performing image compression. ANN-based compression schemes have outperformed conventional handengineered algorithms for lossy and lossless image compression. We have designed an ad-hoc ANN-based image compression scheme to reduce the amount of data needed to be stored and retrieved on space missions studying solar dynamics. In this work, we propose an attention module to make use of both local and non-local attention mechanisms in an adversarially trained neural image compression network. We have also demonstrated the superior perceptual quality of this neural image compressor. Our proposed algorithm for compressing images downloaded from the SDO spacecraft performs better in rate-distortion tradeoff than the popular currently-in-use image compression codecs such as JPEG and JPEG2000. In addition we have shown that the proposed method outperforms state-of-the art lossy transform coding compression codec, i.e., BPG.

AstroDLLC: Efficiently Reducing Storage and Transmission Costs for Massive Solar Observation Data via Deep Learning-based Lossless Compression

“…Other related references on image compression include (Abd-Alzhra & Al-Tamimi 2022; Mishra et al 2022). Zafari et al (2023bZafari et al ( , 2023a proposed a transformer-based multispectral neural image compressor, which uses the VAE model to perform lossy compression on astronomical images, resulting in better rate-distortion performance. Kingma et al (2019) presented a recursive bits-back coding technique called Bit-swap for lossless compression with hierarchical latent variables.…”