Hao Feng

Hi, we use the code "bm = (2 * (bm / 286.8) - 1) * 0.99" rather than "bm = (2 * (bm / 287) - 1)". The goal is...

@lsabrinax It is not used to calculate the flow loss.

@lsabrinax Yeah~

Hi, there is a "docImg" folder in [DRIC dataset](https://hkustconnect-my.sharepoint.com/personal/xliea_connect_ust_hk/_layouts/15/onedrive.aspx?id=%2Fpersonal%2Fxliea%5Fconnect%5Fust%5Fhk%2FDocuments%2FdocDataset&ga=1).

 For me, I can see this page~

Hi. Thanks for your reply! I can not get access to the link, as follows, 

I have submitted your application just now. Thank you!

@an1018 Hi, please use the [OCR eval code](https://github.com/fh2019ustc/DocTr/blob/main/OCR_eval.py) in our repo, in which we have updated the image list used in the [DewarpNet](https://github.com/cvlab-stonybrook/DewarpNet/blob/master/eval/ocr_eval/ocr_files.txt). Then you can obtain the performance as...

@an1018 For more OCR performance of other methods under the two settings (DocTr and DewarpNet), you can refer to the [DocScanner](https://arxiv.org/pdf/2110.14968v2.pdf).

@an1018 Hope to get your reply.