kensakurada / sscdnet Goto Github PK

Hello! Thank you for your job!
Could you please tell me where can I find requirements.txt?
Thank you!

First of all, thank you for your hard work on this and for sharing the training data.

Can you please explain the cvset argument? I see it is calling different training data sets and that it involves five-fold cross-validation. I'm just not sure how you are implementing it.

Hi, I'm interested in the performance of SSCDNet.

Would you please share the pretrained SSCDNet model, or just the pcd_5cv training dataset described in README? I just want to train this model by myself and test its effect.

Thanks.

TSUNAMI and GSV in Panoramic Change Detection dataset are available through an e-mail contact described here including the dataset used for five-fold cross validation in our paper, in which image cropping and data augumentation have been performed.

I click here，but it show that the page is not found

Hi

what are the 5 sets of pscd dataset, can you please release the dataset in the same format as you have shown in the readme that is different CV sets. Also do you have a script to preprocess the pcd dataset.
thanks

Hi everyone,
I try to run this repository by following read.me but I have ModuleNotFoundError: No module named 'correlation_cuda' error. Is there anyone who can get the same error and solve it? Or any recommendation for possible solutions of this errors. Thanks in advance.

Hi, thank you for the sharing of your work, I'm wondering if it's possible to release your checkpoints in your work as well?

usage: train.py [-h] --checkpointdir CHECKPOINTDIR --datadir DATADIR
[--use-corr] [--max-iteration MAX_ITERATION]
[--num-workers NUM_WORKERS] [--batch-size BATCH_SIZE]
[--cvset CVSET] [--icount-plot ICOUNT_PLOT]
[--icount-save ICOUNT_SAVE]
train.py: error: the following arguments are required: --checkpointdir, --datadir

Hi,

Is there any update on the dataset?

Thank you.

In the folder sscnet-master, there is the data set pcd_ 5cv

Thank you for your job and sharing the dataset!
Could you please tell me the more details about the PSCD dataset?
For example, I am not sure what the ground truth used in the training of CSSCDNet mean. Is them showing the change semantic labels from one image to the another in the image pair? Why the two ground truths are not complementary, in some sense.