search

herobd / handwriting_line_generation

Code for BMVC2020 paper "Text and Style Conditioned GAN for Generation of Offline Handwriting Lines"
Other
66 stars 28 forks source link

IAM Dataset #5

Open nuges01 opened 3 years ago

nuges01 commented 3 years ago

This looks like really cool work! My issue isn't with the code/functionality as it is about get a hold of one of the datasets, but I was hoping you could help. I assume I'd need the IAM or DIMES dataset to replicate the results in the paper, but I think the source of the IAM dataset is no longer active. I signed up but didn't get the promised activation link, and the website homepage says "The activities of the Research Group on Computer Vision and Artificial Intelligence terminated with the retirement of its head, Prof. Horst Bunke, by July 31, 2011."

This is curious to me since a lot of (recent) work in handwriting synthesis uses the IAM dataset and a cursory search doesn't reveal any other way to download the dataset. Do you have any insights?

Thanks!

herobd commented 3 years ago

I believe someone is still maintaining the IAM dataset stuff, but they take little a while. How long has it been?

nuges01 commented 3 years ago

Ah. I thought it was automated. I put in the request 2 days ago. I guess I'll be patient.

Is there any pre-processing required for the dataset, or does it work out of the box? It appears that you code uses the form images and the xml files?

herobd commented 3 years ago

The dataset comes with the form, line, and word images. The supplied line images (and word images?) have a lot of the background croped out (white, instead of paper texture), so my code crops the lines from the form images using the xml. My code should run on it as is (the data directory in the repo has the split and character files my code uses).

nuges01 commented 3 years ago

Still no response. Has anyone had any luck getting a response from them recently?

herobd commented 3 years ago

I don't know of anyone. Have you tried using their contact form? (https://fki.tic.heia-fr.ch/contact-info) If you don't get a response there, I'd trying contacting some previous members of the group to see if they know who is supposed to be maintaining it. And let me know what happens, this is a very frequently used dataset in handwriting recognition.

Some former members who I'd expect to be helpful: Marcus Liwicki: marcus.liwicki@ltu.se Andreas Fischer: andreas.fischer@unifr.ch Also you could try an author of the original dataset paper, Urs-Viktor Marti: urs-viktor.marti@swiss.com (?)