Closed Matlmr closed 6 years ago
Please read Dataset under section 4.1 from this paper. They cite a paper titled "Semantic contours from inverse detectors". You can get the full dataset from that paper's website. Let me know if you have any issue searching for the dataset. Thanks!
Sorry for the late answer. On the website you indicated, the ground truth (cls and inst) is availabel as .mat files. How should I convert them to images? Also, it seems to me that it isn't the augmented dataset that corresponds to the train_aug.txt in your repository, any way to get this one?
I did not augment the data, it was available as it is from the paper's website I told you about. But, files were present as .mat files and I had to extract the image and gt from them. I have uploaded my train data, you may download from here (please request access if it is not shared with you, I will share with you).
I requested access. Thanks a lot!
I gave you access.
Hi, great code!
Could you specify exactly which dataset corresponds to train_aug.txt and val.txt and where to get it ? It is the augmented VOC2012, right?