Qiang, Harry; Kounkel, Marina; Bass, Sally; Lingg, Ryan; Sizemore, Logan; Carroll, Dylan; Hutchinson, Brian; Stassun, Keivan G. 鈥�.鈥� Astrophysical Journal 984, no. 1 (2025): 49.听.听
A new technique is introduced for identifying听variable stars鈥攕tars that change in brightness over time鈥攗sing not just simple light curves but a听spatiotemporal data cube, which captures how brightness varies across both space and time. This method allows for better handling of background noise and makes it possible to detect variability across the entire听point-spread function. Such an approach is especially useful for large astronomical surveys like the听Transiting Exoplanet Survey Satellite (TESS).听
The technique was applied to TESS full-frame images to search for听eclipsing,听pulsating, and听rotating听variable stars. A听neural network, trained on known variable stars from other surveys, was used for classification. Approximately听3.1 million candidate variable sources听were identified. While these candidates require further validation using traditional methods, compiling a catalog of the most promising ones can greatly reduce the number of light curves that need to be manually examined.听
As future sky surveys such as the听Legacy Survey of Space and Time (LSST)听begin large-scale photometric monitoring, this type of analysis鈥攚ith improved training data鈥攃ould offer an efficient and scalable way to classify variable stars.听
Figure 1.听Example of phase-folded TESS light curves extracted with eleanor (A. D. Feinstein et al. ) of different types of variables in the training set.