PhD thesis supervisor: dr. Vladas Vansevičius (apply for recommendation)
Detection and cassification of astronomical objects in sky surveys
With the rapidly expanding capabilities of sky observations, the problem of automatic recognition of astronomical objects in large sky surveys has become critical. So far, there is no universal solution to this problem; usually it is solved by engaging numerous volunteers to search for a predetermined class of objects in sky photos interactively (https://www.zooniverse.org/projects?discipline=astronomy).
Relevance. A number of modern ground-based and space-based observatories will begin operating this decade, further emphasizing the importance of solving the problem of automatic identifying astronomical objects.
Novelty. For the first time, the methods of artificial neural network (ANN) and photometric parameter determination will be combined with expert capabilities of the astronomical object recognition.
Perspective. So far, efforts to solve this problem using only ANN methods have not yielded the expected results. By combining ANN image recognition and precision astrophotometry with human cognitive abilities in image analysis, a method suitable for recognizing images of astronomical objects with irregular morphological structures will be developed. The method will be tested on star clusters in nearby galaxies.