semisup-semseg
semisup-semseg copied to clipboard
sud0301
Thank you very much for your work. Here are the problems I encountered in the process of reproduction. The results of my experiment with s4GAN are much better than those...
Thank you for your excellent work and open source code. Ask how to set up semi-supervised experiments with weak labels.
Is the role of MLMT to filter out non-existent prediction classes? Is there no effect on the prediction error between the existing classes?
Thanks for ur wonderful contributions! I wonder how to install the Detail package in pcontext_loader.py line 28? Looking forward to your reply soon!
Kindly tell how to decide whether to use 250 or 255 as in train ids there are no 250 defined train ids in cityscapes dataset according to https://github.com/mcordts/cityscapesScripts/blob/master/cityscapesscripts/helpers/labels.py 
Hi authors, Thank you very much for sharing the code! In the paper, consistency loss is used for all available samples, but Why you only compute unlabeled data in the...
Hello, thank you for your outstanding research work! I encountered such a problem. When I ran train_s4GAN.py (labeled-ratio 0.125/PASCAL-VOC Dataset/number_steps 35000), the best IOU I got was about 0.65, which...
if args.save_output_images: if args.dataset == 'pascal_voc': filename = os.path.join(args.save_dir, '{}.png'.format(name[0])) color_file = Image.fromarray(colorize(output).transpose(1, 2, 0), 'RGB') color_file.save(filename) elif args.dataset == 'pascal_context': filename = os.path.join(args.save_dir, filename[0]) scipy.misc.imsave(filename, gt) why the results...
Hi authors, In line 73 of voc_dataset.py, you use `image = image[:, :, ::-1]` to convert images from BGR to RGB, but in line 192 (TestSet), you don't convert the...
Thanks for your code. We cannot reproduce the results with DeepLab-V3+ by following the default configuration. We have tried using ImageNet pre-trained ResNet-101 as the backbone to implement the DeepLab-V3+...
