Image-based Classification of Variable Stars: First Results from Optical Gravitational Lensing Experiment Data

TL;DR

This paper demonstrates the use of convolutional neural networks to classify variable stars from optical survey images, achieving high accuracy and offering a fast, automated alternative to traditional light curve analysis methods.

Contribution

It introduces an image-based CNN approach for classifying variable stars using phase-folded light curve images, improving automation and accuracy over traditional methods.

Findings

Achieved 80-99% accuracy on OGLE-III data

Achieved 77-98% accuracy on OGLE-IV data

Validated the method's effectiveness across different datasets

Abstract

Recently, machine learning methods presented a viable solution for automated classification of image-based data in various research fields and business applications. Scientists require a fast and reliable solution to be able to handle the always growing enormous amount of data in astronomy. However, so far astronomers have been mainly classifying variable star light curves based on various pre-computed statistics and light curve parameters. In this work we use an image-based Convolutional Neural Network to classify the different types of variable stars. We used images of phase-folded light curves from the OGLE-III survey for training, validating and testing and used OGLE-IV survey as an independent data set for testing. After the training phase, our neural network was able to classify the different types between 80 and 99%, and 77-98% accuracy for OGLE-III and OGLE-IV, respectively.