I posted this to news:
http://www.diybookscanner.org/news/?p=102
Progress on dewarping functionality
Moderator: peterZ
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Progress on dewarping functionality
Another progress update video.
I've finished the manual dewarping and demonstrating some promising work in progress on the automated one.
I've finished the manual dewarping and demonstrating some promising work in progress on the automated one.
Scan Tailor experimental doesn't output 96 DPI images. It's just what your software shows when DPI information is missing. Usually what you get is input DPI times the resolution enhancement factor.
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Progress on dewarping functionality
I blogged it and donated some beer money. Thank you, Tulon! This is really great work.
Re: Progress on dewarping functionality
Thks for the video and your work Tulon.
- rob
- Posts: 773
- Joined: 03 Jun 2009, 13:50
- E-book readers owned: iRex iLiad, Kindle 2
- Number of books owned: 4000
- Country: United States
- Location: Maryland, United States
- Contact:
Re: Progress on dewarping functionality
Wow, Tulon, that is some impressive work! I really like the debug feature!
--Rob
--Rob
The Singularity is Near. ~ http://halfbakedmaker.org ~ Follow me as I build the world's first all-mechanical steam-powered computer.
Re: Progress on dewarping functionality
Really impressive work, Tulon! The progress you're showing in the video is just incredible.
The opinions expressed in this post are my own and do not necessarily represent those of the Canadian Museum for Human Rights.
Re: Progress on dewarping functionality
Thanks guys.
BTW, I will probably need your help on the maths front soon. I am going to try to approximate the text lines with X-Splines. I already figured out how to do it if you can pre-assign tension factors to each control point and t values for each sample. I used this paper for inspiration. It's about B-splines, but it's straightforward to apply it to an X-Spline with hardcoded tension factors. It's already implemented in XSpline::fit() [warning: severe lack of documentation] and there is a commented out example somewhere in the source code. Now, while hardcoding tension factors is fine, pre-assigning t values is generally not. If I don't, I believe the problem will become a non-linear least squares one. That's something beyond my maths skills, but I know you can do it. Take a look at this paper as well. It may be an overkill for our purposes, but who knows. It's about B-splines again, but might be applicable to X-splines.
That would be great, Rob! The code is in Git. Feel free to hack on it and ask questions.rob wrote:I really like the debug feature!
BTW, I will probably need your help on the maths front soon. I am going to try to approximate the text lines with X-Splines. I already figured out how to do it if you can pre-assign tension factors to each control point and t values for each sample. I used this paper for inspiration. It's about B-splines, but it's straightforward to apply it to an X-Spline with hardcoded tension factors. It's already implemented in XSpline::fit() [warning: severe lack of documentation] and there is a commented out example somewhere in the source code. Now, while hardcoding tension factors is fine, pre-assigning t values is generally not. If I don't, I believe the problem will become a non-linear least squares one. That's something beyond my maths skills, but I know you can do it. Take a look at this paper as well. It may be an overkill for our purposes, but who knows. It's about B-splines again, but might be applicable to X-splines.
Scan Tailor experimental doesn't output 96 DPI images. It's just what your software shows when DPI information is missing. Usually what you get is input DPI times the resolution enhancement factor.
Re: Progress on dewarping functionality
I too made a donation. I consider it a pleasure to be able to show a token a gratitude for the effort you are putting into this project. Looking forward to the automatized dewarping!
-
- Posts: 3
- Joined: 08 Sep 2010, 15:40
Re: Progress on dewarping functionality
The test build page only shows a link to a .torrent file. But there are no seeders. Can anyone post another link to this build? I'd really like to try out the new functions.Tulon wrote:So, I've made another progress update video and a test build you may try out.
Re: Progress on dewarping functionality
Scan Tailor experimental doesn't output 96 DPI images. It's just what your software shows when DPI information is missing. Usually what you get is input DPI times the resolution enhancement factor.