dependencies
所属分类:matlab编程
开发工具:matlab
文件大小:25917KB
下载次数:11
上传日期:2018-01-16 20:02:38
上 传 者:
nicaise
说明: attribut glcm for color imaging
文件列表:
dependencies\CCRColour.m (3772, 2011-10-26)
dependencies\ClusterHistogram.m (597, 2011-10-26)
dependencies\codeLBP3x3.m (1217, 2011-10-26)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_HSV_27.mat (205862, 2008-04-12)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_HSV_64.mat (223296, 2008-04-12)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_HSV_8.mat (200519, 2008-04-12)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_Lab_27.mat (422141, 2008-05-08)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_Lab_64.mat (539592, 2008-05-08)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_Lab_8.mat (306541, 2008-05-08)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_RGB_27.mat (136766, 2008-04-11)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_RGB_64.mat (143505, 2008-04-12)
dependencies\Colormaps\INDEXED_SPACE_EUC_MAP_RGB_8.mat (124245, 2008-04-11)
dependencies\Colormaps\MAP_HSV_27.mat (311, 2008-04-11)
dependencies\Colormaps\MAP_HSV_64.mat (443, 2008-04-11)
dependencies\Colormaps\MAP_HSV_8.mat (223, 2008-04-11)
dependencies\Colormaps\MAP_Lab_27.mat (473, 2008-05-08)
dependencies\Colormaps\MAP_Lab_64.mat (752, 2008-05-08)
dependencies\Colormaps\MAP_Lab_8.mat (280, 2008-05-08)
dependencies\Colormaps\MAP_RGB_27.mat (218, 2008-04-11)
dependencies\Colormaps\MAP_RGB_64.mat (221, 2008-04-11)
dependencies\Colormaps\MAP_RGB_8.mat (199, 2008-04-11)
dependencies\Colormaps\old\INDEXED_SPACE_EUC_MAP_HSV_27.mat (205450, 2008-04-05)
dependencies\Colormaps\old\INDEXED_SPACE_EUC_MAP_HSV_64.mat (213840, 2008-04-05)
dependencies\Colormaps\old\INDEXED_SPACE_EUC_MAP_HSV_8.mat (200370, 2008-04-05)
dependencies\Colormaps\old\INDEXED_SPACE_EUC_MAP_RGB_27.mat (138706, 2008-04-04)
dependencies\Colormaps\old\INDEXED_SPACE_EUC_MAP_RGB_64.mat (150608, 2008-04-05)
dependencies\Colormaps\old\INDEXED_SPACE_EUC_MAP_RGB_8.mat (124245, 2008-04-04)
dependencies\Colormaps\old\MAP_HSV_27.mat (280, 2008-01-10)
dependencies\Colormaps\old\MAP_HSV_64.mat (406, 2008-01-10)
dependencies\Colormaps\old\MAP_HSV_8.mat (220, 2008-01-10)
dependencies\Colormaps\old\MAP_Lab_27.mat (475, 2008-01-10)
dependencies\Colormaps\old\MAP_Lab_64.mat (768, 2008-01-10)
dependencies\Colormaps\old\MAP_Lab_8.mat (277, 2008-01-10)
dependencies\Colormaps\old\MAP_Luv_27.mat (468, 2008-02-05)
dependencies\Colormaps\old\MAP_Luv_64.mat (743, 2008-02-05)
dependencies\Colormaps\old\MAP_Luv_8.mat (283, 2008-02-05)
dependencies\Colormaps\old\MAP_RGB_27.mat (210, 2008-01-11)
dependencies\Colormaps\old\MAP_RGB_64.mat (224, 2008-01-11)
dependencies\Colormaps\old\MAP_RGB_8.mat (192, 2008-01-11)
dependencies\Colormaps\old\MAP_sRGB_27_.mat (274, 2008-01-11)
... ...
Gabor Filtering Toolbox v 0.1 for Matlab
----------------------------------------
This is the initial released version, some bugs will be likely
found. It has been tested mainly with Matlab R14.
Short introduction to usage of Gabor toolbox
--------------------------------------------
First, an image is loaded and converted to double. The image has
likely three color channels even if it is a gray-scale image, so only
one of them must be selected. While not strictly necessary, the image
can be also scaled to [0,1] instead of normal [0,255].
image=imread('someimage.jpg'); % load image
image=image(:,:,1); % select only one color channel
image=double(image)./256; % convert to double and scale to [0,1]
Then, a filterbank can be created.
bank=sg_createfilterbank(size(image), 0.2 , 5, 4,'verbose',1);
creates a filterbank with the size of the image, the frequency of the
highest frequency filter is 0.2, 5 filters at different frequencies
and 4 orientations are created. The created filter bank will be
displayed. Only half of the filterbank is created, because
responses for the second half of the filter bank are complex
conjugates of the responses from the first half.
The filterbank can be used to filter images.
r=sg_filterwithbank(image,bank);
The responses will be returned in a special structure. The structure
can be converted to a 3-d matrix by using
m=sg_resp2samplematrix(r);
Now, m will be a 512x512x20 (or whatever the image resolution was x20)
matrix, since there are 5*4=20 Gabor filters. If you do not have a
good idea what to do with the responses, you can for example view a
spooky image by summing all the responses:
imagesc(abs(sum(m,3))); colormap(gray);
For object detection and localization functionality other functions
are also available.
Sample matrix can be normalized for illumination invariance:
m_norm=sg_normalizesamplematrix(m)
For scale invariance, extra frequencies can be first included in
the filter bank:
bank=sg_createfilterbank(size(image), 0.2 , 5, 4,'extra_freq',1);
r=sg_filterwithbank(image,bank);
m=sg_resp2samplematrix(r);
Then, different scales of features can be selected from sample
matrix:
m2=sg_scalesamples(m,0,5,4)
will select the 5 highest frequencies and
m2=sg_scalesamples(m,1,5,4)
the lower available 5 frequencies. Note that if normalization is
required, it must be done after this sg_scalesamples step.
For rotation invariance, the sample matrix can be rotated:
m2=sg_rotatesamples(m,1,4)
rotates the responses by 1 orientation.
近期下载者:
相关文件:
收藏者: