search

MarcWong / UDD

Urban Drone Dataset(UDD) for "Large-scale Structure from Motion with Semantic Constraints of Aerial Images", PRCV2018
MIT License
82 stars 9 forks source link

question! #3

Closed zhujiesuper closed 3 years ago

zhujiesuper commented 3 years ago

your job is awesome ,but can you tell me the mean of m1-train200-val40-test48? it is a new dataset?

MarcWong commented 3 years ago

This is just the subdivision of M1 with ground truths. And there is a little overlap between UDDs and M1. If you are training this task, just use UDD5 or UDD6 would be enough. Actually the m1-train200-val40-test48 is discarded somehow, no need to maintain the ground truth of it.

zhujiesuper commented 3 years ago

thank you!

------------------ 原始邮件 ------------------ 发件人: "MarcWong"<notifications@github.com>; 发送时间: 2020年9月5日(星期六) 晚上9:53 收件人: "MarcWong/UDD"<UDD@noreply.github.com>; 抄送: "朱一大爱"<653725215@qq.com>; "Author"<author@noreply.github.com>; 主题: Re: [MarcWong/UDD] question! (#3)

This is just the subdivision of M1 with ground truths. And there is a little overlap between UDDs and M1. If you are training this task, just use UDD5 or UDD6 would be enough. Actually the m1-train200-val40-test48 is discarded somehow, no need to maintain the ground truth of it.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.