DynamicDepth icon indicating copy to clipboard operation
DynamicDepth copied to clipboard

The number of cityscape dataset

Open Qianqian3764 opened this issue 1 year ago • 3 comments

Thank your excellent job! I found that the number of training sets is not the same as the official division. There are a certain amount of cuts (69731 vs 52619) in the article. Will this affect the final results?

Qianqian3764 avatar May 22 '24 16:05 Qianqian3764

Thank your excellent job! I found that the number of training sets is not the same as the official division. There are a certain amount of cuts (69731 vs 52619) in the article. Will this affect the final results?

The same thing happened to me. Did you solve it?

fufj avatar Sep 24 '24 14:09 fufj

Not yet.

Qianqian3764 avatar Oct 04 '24 09:10 Qianqian3764

Thank your excellent job! I found that the number of training sets is not the same as the official division. There are a certain amount of cuts (69731 vs 52619) in the article. Will this affect the final results?

Have you found any segmentation networks or masks for identifying dynamic objects in this article that uses a segmentation network to recognize dynamic objects?

757787182 avatar Jan 16 '25 15:01 757787182