Determination of Metallicity of different Galaxies using Image Processing and DeepLearning Algorithms

DOI

No DOI

Published In

Volume 3, Issue 4 (2022)

Published Online

July-August 2022

Pages

194-199

Article Type

Research Article

Check for updates

No DOI available for cross-reference

Abstract: Using spectroscopic knowledge, gas-phase metallicity was predicted for galaxies by training a deep convolutional neural network (CNN) with the help of fastai built on PyTorch. The three band gri images obtained from Sloan Digital Sky Survey (SDSS) are trained and tested. The deep residual CNN prevails over a random forest algorithm over similar dataset, as the RMSE (Root Mean Square Error) for the CNN is 0.092 dex, whereas the random forest algorithm gives out an RMSE of 0.130 dex.