An Investigation of Machine Learning Techniques for Object Classification in Sequential Astronomical Imagery

This study addresses the problem of classifying objects in a video sequence of the starry sky, specifically the task of distinguishing background stars from moving objects. The main focus is on a comparative analysis of two approaches: the Support Vector Machine (SVM) method and a neural network-based method. For the SVM, object motion trajectories were used as features, while the neural network processes these trajectories and extracts features automatically.