search

nie-lang / DeepRectangling

CVPR2022 (Oral) - Deep Rectangling for Image Stitching: A Learning Baseline
226 stars 37 forks source link

VGG model download #2

Open FXY0117 opened 2 years ago

FXY0117 commented 2 years ago

Excuse me, would you mind providing imagenet-vgg-verydeep-19.mat in Baidu Cloud, I have been downloaded this model in the page but the speed is two slow. Thank you !

nie-lang commented 2 years ago

链接：https://pan.baidu.com/s/1Necpa-b9SJXorRXXCZQC-w 提取码：1234

FXY0117 commented 2 years ago

Excuse me, is there something wrong with your DIR-D dataste in Baidu Cloud link, I can't find the dataset but some files.

nie-lang commented 2 years ago

Thanks for your advice. I have updated the link.

FXY0117 commented 2 years ago

Excuse me, for a stitched image dataset, currently,it is irrectangular, so how can I get the gt file. May I copy the input file and rename gt ? What's more, would you mind providing the code of mask generation in this github?

nie-lang commented 2 years ago

1) I have updated the codes for images with arbitrary resolution, and this version requires no gt files. You can try it. 2) In fact, I get the masks by the stitching algorithms. For example, two images are stitched by mesh deformation. And I use the same mesh warp to stitch two white images to generate the perfect mask. Or you can get the mask via threshold segmentation, the masks with some noise are also OK.