CRAFT-Reimplementation

CRAFT-Reimplementation copied to clipboard

Bug report: random crop

7

If the thread goes into [this](https://github.com/backtime92/CRAFT-Reimplementation/blob/master/data_loader.py#L93) branch, incorrect images like below occur:  If it goes to the `else `branch, the results are fine:  I think this is because...

ThisIsIsaac

question on wechat

1

I want to join the WeChat group, but the QR code has expired. Can I resend it or pull me into the group?

zcswdt

how to improve the recall? when I trained on IC2015, the precision is up to 0.85, but the recall is always in 0.65-0.72?

2

yaya325

Questions on scaling images and GT masks

1

1. In `data_loader.py` , `pull_item` function: ```python region_scores = self.resizeGt(region_scores) affinity_scores = self.resizeGt(affinity_scores) confidence_mask = self.resizeGt(confidence_mask) ``` and the function definition of `resizeGt` is: ```python def resizeGt(self, gtmask): return cv2.resize(gtmask,...

ThisIsIsaac

Question on Maploss

2

While reading your loss function, I came across a few questions: 1. I think `def single_image_loss` defined in `mseloss.py` seems to be a misnomer. The function iterates through outputs for...

ThisIsIsaac

fixes [issue](https://github.com/backtime92/CRAFT-Reimplementation/issues/28)

ThisIsIsaac

when and when not to freeze

5

You freeze `vgg16_bn` weights in `trainic15data.py` but not in `trainSyndata.py`. 1. Why do you freeze in `trainic15data.py` and not in `trainSyndata.py`? 2. What is the benefit of freezing? I assumed...

ThisIsIsaac

Have you sorted boundry boxes?

2

nbhupendra

Critical issues with training code

2

1. training scripts (`trainSyndata.py`) imports `test.py`. `test.py` has some code not nested in `if __name__ == "__main__":`. For example, this below code that parses argument: ```python parser = argparse.ArgumentParser(description='CRAFT Text...

ThisIsIsaac

about loss function

1

i am confused about loss function in the paper, when estimated character bounding boxes is low mean the accuracy of model is bad, confidence score is low and the loss...

cuongdxk57