How to improve the extracted image quantity

[LeighSu]

when to use the marker_single 02-FEDERATED\ LEARNING\ FOR\ KEYWORD\ SPOTTING.pdf test, it can extract the correct image. but I find the picture is blurry. is there any option or method to improve the extracted image quantity to make it more clear.

ps:the pdf link is https://arxiv.org/abs/1810.05512

[VikParuchuri]

We will add an option for this in an upcoming release