Criminisi
所属分类:图形图像处理
开发工具:matlab
文件大小:340KB
下载次数:34
上传日期:2017-06-26 15:26:49
上 传 者:
Shineas
说明: criminisi图像修复程序 修复破损图
(criminisi to repair the image ,that is very good)
文件列表:
20.png (162139, 2015-03-26)
201.png (136031, 2015-03-26)
benggee_7.jpg (37938, 2015-03-29)
bestexemplarhelper.asv (2382, 2015-04-06)
bestexemplarhelper.c (2503, 2015-04-30)
bestexemplarhelper.mexmaci64 (8536, 2013-03-12)
bestexemplarhelper.mexw32 (6656, 2015-03-26)
Evaluate.m (293, 2015-04-30)
inpaint7.asv (8928, 2015-05-03)
inpaint7.m (8928, 2015-05-03)
plotall.m (486, 2015-04-11)
CS7495 Final Project
Sooraj Bhat
Object Removal by Exemplar-based Inpainting
USING THE CODE
Everything was done in Matlab and MEX (i.e. a C function callable from
Matlab). First, the C code needs to be compiled. At the Matlab
prompt, type:
mex bestexemplarhelper.c
You should then be able to run the following (which takes ~2mins on my
2GHz Pentium 4 laptop with 256MB of RAM):
[i1,i2,i3,c,d,mov] = inpaint('bungee0.png','bungee1.png',[0 255 0]);
A smaller run (~25 seconds) would be:
[i1,i2,i3,c,d,mov] = inpaint('bw0.png','bw2.png',[0 255 0]);
To visualize:
plotall;
close; movie(mov);
SAMPLE IMAGES & RESULTS
*0.png The original images ('bw' and 'bungee')
*{1,2}.png The different fill region masks.
*P2.png Plots of confidence and data terms.
*P5.png Plots of original, fill region, inpainted images and
confidence and data terms.
Other images in the dir contain redundant information. I didn't
include the movies since they are big, but they are available on my
project page (AVI format).
近期下载者:
相关文件:
收藏者: