adrita-khan / astro_u-net

Astro_U-net

Astro-U-net: Learning to Enhance Astronomical Images

Astronomical images are essential for exploring and understanding the universe. Images often contains additive noise, which makes denoising mandatory step in post-processing data before further data analysis. In order to maximise the efficiency and information gain in the post-processing of astronomical imaging, we decided to turn on machine learning. We propose Astro U-net, a fully-convolution neural network for image denoising and enhancement. We made many experiments and picked two of the best performing networks -- Network 1 and Network 2. More details are provided in the paper.

To watch ESAC TechTalk click here :).