auto_thumb
所属分类：模式识别(视觉/语音等)
开发工具：matlab
文件大小：57KB
下载次数：40
上传日期：2010-05-19 20:03:32
上 传 者：
xiaobf
说明： 这是2003年有关图像自动裁剪的源码，在人脸识别的预处理中非常有用，来自国外网站
(This is 2003, the image automatically cut the source, pre-processing in face recognition is very useful site from abroad)
文件列表:
auto_thumb (0, 2010-05-14)
auto_thumb\AS0100.jpg (35795, 2003-02-21)
auto_thumb\auto_thumbnail.m (2016, 2004-01-12)
auto_thumb\CompColorConspicuity.m (2228, 2004-01-10)
auto_thumb\CompIntensityConspicuity.m (1712, 2003-02-14)
auto_thumb\CompOrientationConspicuity.m (3455, 2003-02-20)
auto_thumb\CompSaliencyMap.m (1004, 2003-02-14)
auto_thumb\compuAuxSum.c (4480, 2003-01-29)
auto_thumb\Cropping.m (1814, 2003-03-07)
auto_thumb\cropping_BF.c (9472, 2003-02-05)
auto_thumb\cropping_BF.dll (8704, 2003-02-05)
auto_thumb\Cropping_Heuristic.m (4771, 2003-06-22)
auto_thumb\DoGFilter.m (1149, 2003-01-20)
auto_thumb\FindThres.m (3543, 2004-06-17)
auto_thumb\gaborconvolve.m (6490, 2004-01-12)
auto_thumb\GaussianPyramid.m (1740, 2003-02-22)
auto_thumb\PostProcesSaliencyMap.m (1517, 2003-03-09)
auto_thumb\Scale.m (342, 2003-02-21)
auto_thumb\SurroundInhibit.m (1785, 2003-03-09)
auto_thumb\ThumbCrop.m (1535, 2003-01-19)
auto_thumb\ThumbCrop_Demo.m (2229, 2003-01-22)
auto_thumb\thumb_crop.m (4117, 2004-08-03)
The function 'auto_thumbnail' in auto_thumbnail.m is the main function here.
The function is of the following form:
[imout] = auto_thumbnail(imin, iFigure)
It takes imin as the input image and returns imout as the cropped image
which is to be shrinked to a thumbnail.
If iFigure>0, then demo result in the figure with id iFigure, iFigure can also be ignored.
The basic idea is to first compute the saliency map of the given image, then cropping
the image to contain as more saliency value as possible while with as small size as possible.
The idea can be refer to paper
B. Suh, H. Ling, B.B. Bederson, D.W. Jacobs, Automatic Thumbnail Cropping and Its Effectiveness, ACM UIST 2003.
All the code is written by Haibin Ling, except that Gabor convolution part is modified from code by
Peter Kovesi. Any comments are welcome to hbling AT umiacs.umd.edu .
近期下载者：
相关文件：
收藏者：