I've prepared a dataset of 2726 meteors and 1131 false detections, published here
. I'm trying to use machine learning to do binary classification of similar images: whether they show a meteor or not. My experience with machine learning (in Python) is limited, however, so I decided to publish the dataset for anyone who might find it interesting, and perhaps I can get some help in making a good classifier. My initial attempt (included in the dataset) was moderately promising, but I believe it can be improved.
Anyone please feel free to have a go at this if it sounds interesting. Read more: Machine learning challenge
Examples of actual meteors from the dataset:
Examples of false detections: