Wenjun Wang

one question about SR

8

If in the original image, one pixel represents the actual distance of 2mm, after 4x super-resolution, will one pixel represent the actual distance of 8mm?

Deeplabv3plus inference time

5

HI, I have a confusion hoping to get help. Why is the inference speed of Deeplabv3plus much lower than other models with similar parameters and computational complexity（GFLOPs）?

not enough values to unpack (expected 2, got 1) height, width = img.shape[:2]

2

def get_training_augmentation(): train_transform = [ A.HorizontalFlip(p=0.5), A.ShiftScaleRotate(scale_limit=0.5, rotate_limit=0, shift_limit=0.1, p=1, border_mode=0), A.PadIfNeeded(min_height=320, min_width=320, always_apply=True, border_mode=0), A.RandomCrop(height=320, width=320, always_apply=True), A.IAAAdditiveGaussianNoise(p=0.2), A.IAAPerspective(p=0.5), A.OneOf( [ A.CLAHE(p=1), A.RandomBrightness(p=1), A.RandomGamma(p=1), ], p=0.9, ), A.OneOf( [...

Transformer backbone for dense prediction

11

When using timm, the convolutional neural network can be used as the backbone（features_only=True, out_indices=(1, 2, 3, 4)） for dense prediction, but how can the transformer model be used directly as...

About network

1

I notice that almost all the semi-supervised segmentation models used resnet-deeplab as the network, any special to be considered? will it be better by using transformer network?

Question about learning rate

3

In the paper, you said "The learning rate is set initially to 10−4 for both generator and critic. After the first 150 epochs we linearly decay the rate to zero...

Image Motion Blur code

2

I want to train on my own dataset, but i do not know how to generate blur images for training. Can you provide the code for generating motion blur images?...

model performance

1

Hi,I have a confusion hope can be helped， In semi-supervised segmentation, why the performance improvement always becomes smaller as the amount of labeled data increases?  Thanks!

Thanks for your excellent work, can you share the robutness test code? As shown in the paper"we test 3 severities for 4 kinds of“Noise” and 5 severities for the rest...

question

1

Can i run it in windows10 environment? Thanks