search

PRIS-CV / DSACA

Code release for Dilated-Scale-Aware Category-Attention ConvNet for Multi-Class Object Counting
MIT License
20 stars 1 forks source link

Test code #1

Open JuiceTao opened 6 months ago

JuiceTao commented 6 months ago

Hello,where is the test code?

xuweigogogo commented 6 months ago

Hi, the validation phase is combined into the training code (RSOC_train_class2_CAM_DSAM.py, VisDrone_train_class8_CAM_DSAM.py).

https://github.com/PRIS-CV/DSACA/blob/6dd16208e134a4a78c554de73d6612c77f69bb4c/VisDrone_train_class8_CAM_DSAM.py#L162

llzhuzw commented 2 weeks ago

Hello,I couldn't find the part where you organized the RSOC dataset, so I organized a version by myself. But I'm not sure if it's correct. Could you please check if the number of images in each phase is correct?Thanks!

RSOC_2
xuweigogogo commented 2 weeks ago

Hello,I couldn't find the part where you organized the RSOC dataset, so I organized a version by myself. But I'm not sure if it's correct. Could you please check if the number of images in each phase is correct?Thanks! RSOC_2

Thanks for your attention! However, for some reason (as time passed), I can not analyze the original organized RSOC dataset right now. How about sending the organized dataset to my email? I would check this dataset by looking through the images contained in it.

xuwei2838@foxmail.com

llzhuzw commented 2 weeks ago

Hello,I couldn't find the part where you organized the RSOC dataset, so I organized a version by myself. But I'm not sure if it's correct. Could you please check if the number of images in each phase is correct?Thanks! RSOC_2

Thanks for your attention! However, for some reason (as time passed), I can not analyze the original organized RSOC dataset right now. How about sending the organized dataset to my email? I would check this dataset by looking through the images contained in it.

xuwei2838@foxmail.com

OK

xuweigogogo commented 2 weeks ago

Hello,I couldn't find the part where you organized the RSOC dataset, so I organized a version by myself. But I'm not sure if it's correct. Could you please check if the number of images in each phase is correct?Thanks! RSOC_2

Thanks for your attention! However, for some reason (as time passed), I can not analyze the original organized RSOC dataset right now. How about sending the organized dataset to my email? I would check this dataset by looking through the images contained in it. xuwei2838@foxmail.com

OK

After looking through the dataset you provided, I think you have employed the correct procedures to organize it. The essential procedures are two steps: Step 1: combine the small- and large-vehicle images into a single set; Step 2: combine the annotations of small and large vehicles in the same image.

In my opinion, the dataset you provided is correct. You can check if there are missed details for further confirmation.

Thanks for your efforts.