Oval detection

[lemonci]

You mentioned in your paper: "deep learning based methods can provide image proposals which contain oval objects for image pre-processing while they are still inappropriate to directly handle ellipse detection due to the issues of limited segmentation accuracy and expensive manual annotation" I would like to know which methods you referred to and if they are unsupervised, as my samples contains a lot of oval patterns which may not work well under eclipse detection. Thanks.