Computer vision as a powerful tool for transiting exoplanet search

Abstract

Computer vision is a field of artificial intelligence that deals with how computers can gain understanding from digital images or videos. From another perspective, it seeks to understand and automate tasks that the human visual system can do. Computer vision tools are already in use for galaxy search and classification, or comet search. Finding transiting planet candidates looking at the photometric light curves is easy for humans, therefore they should be detectable with computer vision methods. However, first, the time-series data must be somehow encoded into 2D images. I present a pilot study of transiting planet search with computer vision using Convolutional Neural Network and its application on TESS Sector 1 short cadence light curves.