-
Thank you for your great work! I have already read #13 , but I cannot convert DeepGlobe to .csv for APLS. Could you share your solution? It will be of great help.
-
比赛网站上下不到数据集。。。
-
No such file or directory: '/data/xx/Deepglobe/1/test/groundtruth/36_mask_44.png'
-
I download the trained NL34_LinkNet weight on dropbox. It can be loaded successfully, but there is nothing being draw(Almost black pic), even put in trainset of deepglobe. I don't know what's wrong wi…
-
用作者提供的预训练权重,能测下deepglobe数据集中的27127.jpg这张影像的结果吗,我测试了下,得到的没有论文中的好
-
Thank you very much for your work! I trained the deepglobe dataset with your train_mtl.py , but the prediction is not very good, even worse than using segmentation network like unet or linknet. Could …
-
感谢您的工作!
在拜读您论文时,有两个问题不太明白,特向您请教:
一、请问使用AR工具标注Bounding Box的作用是什么?
我是这样理解的:
1. SFM重建的点云是相机坐标系下的,使用BBox将其转换到目标坐标系下?
2. 多次采集的图像序列重建的点云有偏差,使用BBox将它们融合?
如果第1个理解正确的话,那么使用无AR功能的相机…
-
I'd like to know more about the dataset that was utilized in the research. Can you tell me about it?
-
Hi ,
I have doubt regarding number of classes for mIou , as per paper there are 5 classes Imp surface , building , Low veg , car , Tree and mIou is considered for these 5 classes ,
But in Code ,…
-
- [x] ResUnet: https://medium.com/@nishanksingla/unet-with-resblock-for-semantic-segmentation-dd1766b4ff66
- [x] D-LinkNet: https://github.com/zlkanata/DeepGlobe-Road-Extraction-Challenge/blob/master…