Open zjk958162734 opened 3 years ago
Please see the link here: https://github.com/NVIDIA/semantic-segmentation#downloadprepare-data
I guess you may have misunderstood what I mean. I have downloaded the data set, but I found this kind of image in it.
What I want to ask is, how can I get the following RGB image
Because I want to use this image to train my own model, thank you
------------------ 原始邮件 ------------------ 发件人: "NVIDIA/semantic-segmentation" <notifications@github.com>; 发送时间: 2020年11月30日(星期一) 晚上11:56 收件人: "NVIDIA/semantic-segmentation"<semantic-segmentation@noreply.github.com>; 抄送: "帅气| 宝贝"<958162734@qq.com>;"Author"<author@noreply.github.com>; 主题: Re: [NVIDIA/semantic-segmentation] get auto-label image (#101)
Please see the link here: https://github.com/NVIDIA/semantic-segmentation#downloadprepare-data
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.
If you are asking for the corresponding training RGB input images, you have to download the cityscapes coarse dataset.
https://github.com/NVIDIA/semantic-segmentation/blob/main/PREPARE_DATASETS.md
You have to download the cityscapes train_extra dataset. Please see the link and set up the structure of the directory as shown in the link above.
Hello, I am sorry to interrupt you. I want to ask, if I just want to get the data generated by your automatic tagging, like the third column of figure4 in your paper(Auto-generated coarse label), what should I do, thank you.