prgumd / salientdso Goto Github PK

Reading Photometric Calibration from file /home/kesse/dataset/sequence_01/pcalib.txt
Reading Vignette Image from /home/kesse/dataset/sequence_01/vignette.png
Successfully read photometric calibration!
got 4757 images and 4757 timestamps and 4757 exposures.!
ImageFolderReader: got 4757 files in /home/kesse/dataset/sequence_01/images.zip!
using pyramid levels 0 to 3. coarsest resolution: 80 x 60!
dso_dataset: /home/kesse/SalientDSO/src/main_dso_pangolin.cpp:482: int main(int, char**): Assertion `saliency != ""' failed.

OBS: It's getting this error. what can i do?

Hi, I have read your papers recently, and I have a question, how to set the Object Class‘s weight in the Salience Map in the experiment？

Firstly, I have run SALGAN successfully，however, if I run a dataset from https://vision.in.tum.de/mono-dataset, I only predict it and no groud truth.
Secondly, there's no evaluation procedure for the results of saliency map.
Could you tell me how to get ground truth for the dataset and evaluation procedure? Thank you very much.

Hi, I want to ask a question about "5.3. Test Semantic Filtered Saliency Setting".
I can't solve the problem of how to predict the output to many images (a dataset) once by PSPNet.
so can you help me?
thanks ～～～

hello, do you have the ground truth of camera trajectory about the CVL datasets. I want to compare with the result of salientdso.
thank you very much~~