文件列表:

extractFeatures (0, 2008-11-28)
extractFeatures\extractFeatures.vcproj (4435, 2008-11-21)
extractFeatures\felib.dll (33280, 2008-11-21)
extractFeatures\libfe.0.1.a (39512, 2008-11-22)
extractFeatures\extractFeatures.sln (1833, 2008-11-21)
extractFeatures\extractFeatures.suo (57344, 2008-12-31)
extractFeatures\extractFeatures.vcproj.t61.jackie.user (1426, 2008-12-31)
extractFeatures\libfe-lin.0.1.so (75645, 2008-11-28)
extractFeatures\extractGabor (9695, 2008-11-28)
extractFeatures\readList.m (346, 2008-03-15)
extractFeatures\readbin.m (530, 2008-11-21)
extractFeatures\extractFeatures (14209, 2008-11-28)
extractFeatures\data.bin (245776, 2008-11-28)
extractFeatures\main.map (0, 2008-08-29)
extractFeatures\mexFeatures (0, 2008-08-17)
extractFeatures\mexFeatures\mexFeatures.vcproj (4553, 2008-08-20)
extractFeatures\mexFeatures\mexFeatures.vcproj.t61.jackie.user (1390, 2008-12-31)
extractFeatures\mexFeatures\features.def (46, 2008-08-18)
extractFeatures\include (0, 2008-11-28)
extractFeatures\include\gabor.h (2704, 2008-08-20)
extractFeatures\include\lbp.h (2428, 2008-08-20)
extractFeatures\include\util.h (4681, 2008-09-10)
extractFeatures\include\colormoment.h (2777, 2008-08-20)
extractFeatures\include\edge.h (2577, 2008-08-20)
extractFeatures\include\gist.h (2563, 2008-09-10)
extractFeatures\include\._felib.h (4096, 2008-11-22)
extractFeatures\include\felib.h~ (3676, 2008-11-22)
extractFeatures\include\felib.h (3679, 2008-11-28)
extractFeatures\extractFeatures.ncb (13011968, 2008-12-31)
extractFeatures\data (0, 2008-11-21)
extractFeatures\data\exp.txt (43, 2008-11-21)
extractFeatures\data\apple8.jpg (5479, 2008-09-07)
extractFeatures\data\apple2.jpg (24740, 2008-09-07)
extractFeatures\data\apple6.jpg (3485, 2008-09-07)
extractFeatures\felib.lib (9670, 2008-11-21)
extractFeatures\extractFeatures.exe (9216, 2008-11-21)
extractFeatures\example.cmd (29, 2008-11-21)
extractFeatures\extractGWT (0, 2008-11-21)
extractFeatures\extractGWT\extractGWT.vcproj (4313, 2008-11-21)
... ...

Feature Extraction Library FELib LICENSE This software is freely available for non-commercial use such as research and education. Please see the full disclaimer below. We recommand you cite the reference given below in your publications related to this work. Feature extraction for image retrieval: Jianke Zhu, Steven C.H. Hoi, Michael R. Lyu and Shuicheng Yan,"Near-Duplicate Keyframe Retrieval by Nonrigid Image Matching," ACM Multimedia'2008. Feature extraction for face recognition: Jianke Zhu, Steven C.H. Hoi and Michael R. Lyu,"Face Annotation by Transductive Kernel Fisher Discriminant," IEEE Trans. on Multimedia, vol. 10, pp. 86-96. 2008. Copyright (c) 2003-2009 Jianke Zhu. Email:jianke.zhu at gmail.com http://www.vision.ee.ethz.ch/~zhuji 3RD PART SOFTWARE The software is partly based on the following libraries: - The Intel(tm) OpenCV Library DISCLAIMER This software is provided 'as-is', without any express or implied warranty. In no event will the author be held liable for any damages arising from the use of this software. Permission is granted to anyone to use this software for any non-commercial purpose, and to alter it, subject to the following restrictions: 1. The origin of this software must not be misrepresented; you must not claim that you wrote the original software. 2. Altered source versions must be plainly marked as such, and must not be misrepresented as being the original software. 3. This notice may not be removed or altered from any source distribution. -- No guarantees of performance accompany this software, nor is any responsibility assumed on the part of the author. This software is provided by Jianke Zhu ``as is'' and any express or implied warranties, including, but not limited to, the implied warranties of merchan- tability and fitness for a particular purpose are disclaimed. In no event shall Jianke Zhu be liable for any direct, indirect, incidental, special, exemplary, or consequential damages (including, but not limited to, procurement of substitute goods or services; loss of use, data, or profits; or business interruption) however caused and on any theory of liability, whether in contract, strict liability, or tort (including negligence or otherwise) arising in any way out of the use of this software, even if advised of the possibility of such damage. $Revision: 1.2 $ $Date: 2008/12/31 $ Change log: v1.2 Correct lots of errors in the comiple instructions with help from zim . Add compile.m to facilitate a successful compilation Instruction on nomalizing the vector to zero mean and unit variance. This is important to obtain good retrieval results on a large collection of the diverse images. v1.1 Add GIST feature About the software This software extracts the four kinds of features, 1. Color histogram, color moments 2. Edge histogram 3. Gabor wavelets transform 4. Local Binary Pattern 5. GIST For technical details of the FELib, please refer to the following publication, Jianke Zhu, Steven C.H. Hoi, Michael R. Lyu and Shuicheng Yan, °Near-Duplicate Keyframe Retrieval by Nonrigid Image Matching,± ACM Multimedia'2008. This toolbox is also used in the following publications: Steven C.H. Hoi, Rong Jin, Jianke Zhu, and Michael R. Lyu, "Semi-Supervised SVM Batch Mode Active Learning with Applications to Image Retrieval," To appear in ACM Transactions on Information Systems (TOIS), 2009. Steven C.H. Hoi, Rong Jin, Jianke Zhu, and Michael R. Lyu, "Semi-Supervised SVM Batch Mode Active Learning for Image Retrieval," In Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR2008), Alaska, 24-26 June, 2008. Jianke Zhu, Steven C.H. Hoi and Michael R. Lyu,"Face Annotation by Transductive Kernel Fisher Discriminant," IEEE Trans. on Multimedia, vol. 10, pp. 86-96. 2008. Install Guide A. Install OpenCV 1. Install the OpenCV library: Download opencv from http://sourceforge.net/projects/opencvlibrary, this release support OpenCV 1.0 (Oct. 2006). 2a.Windows Platform Add \lib into MSVC library search path, and \cxcore\include, \cv\include, \otherlib\highgui into MSVC include search path, add \bin into the system path of windows. 2b.Linux Refer to OpenCV FAQ 2c.MAC OSX 10.5 Refer to the guide in the following link: http://opencvlibrary.sourceforge.net/Mac_OS_X_OpenCV_Port **********************************Important Note*************************************************************** To compile it as a stand alone program, you need comment the first line in main.c #define MATLAB_API To obtain a Matlab MEX or .dll, you need ensure the definition of MATLAB_API Also, comment following macro definition to compile under Linux/Mac OSX #define WIN32 **********************************End Note********************************************************************* B. Windows with MSVC compiler: You can open the prject file of MSVC8.0 directly, and build the project. If using prevoius MSVC version, follow the following steps: 1. Create a new project 2. Setup a new empty VC++ console project 3. Add the attached main.c into the project 4. Add cv.lib cxcore.lib highgui.lib (debug version cvd.lib cxcored.lib highguid.lib) into the additional library dependency of link tab 5. Build the project. C. Compile on Linux and Mac OSX It is very easy, just use following command: //extract features in Mac OSX gcc -o extractFeatures -I./include -I/opt/local/include/opencv -L. -L/opt/local/lib -llapack -lblas -lcxcore -lcv -lhighgui -lfe.0.1 main.c //extract features in Ubuntu Linux (Remember to copy 'libfe-lin.0.1.so' into the fold '\usr\lib' or dynamic library searching path, do ldconfig ) gcc -o extractFeatures -I./include -I/usr/include/opencv -L. -L/opt/local/lib -llapack -lblas -lcxcore -lcv -lhighgui -lfe-lin.0.1 main.c //Example: ./extractFeatures data //extract Gabor features in Mac OSX gcc -o extractGabor -I./include -I/opt/local/include/opencv -L. -L/opt/local/lib -llapack -lblas -lcxcore -lcv -lhighgui -lfe.0.1 extractGabor.c //extract Gabor features in Ubuntu Linux (Remember to copy 'libfe-lin.0.1.so' into the fold '\usr\lib' or dynamic library searching path ) gcc -o extractGabor -I./include -I/usr/include/opencv -L. -L/opt/local/lib -llapack -lblas -lcxcore -lcv -lhighgui -lfe-lin.0.1 extractGabor.c //Example: ./extractGabor data D. Matlab Win: % change the mex compiler to VC8. Using LCC, you will get following compiling error % "\cxtypes.h: 300 compiler error in _kids--Bad rule number 0". Thanks zim. mex -setup mex -output mexFeatures -I. -I.\include -IC:\OpenCV\cv\include -IC:\OpenCV\cxcore\include -IC:\OpenCV\otherlibs\highgui -LC:\OpenCV\lib... -L. -lcxcore -lcv -lhighgui -lfelib main.c Linux/ MAC OSX: mex -output mexFeatures -I. -I./include -I/opt/local/include/opencv -L/opt/local/lib -L. -lcxcore -lcv -lhighgui -lfe-lin.0.1 main.c Usage of the library: For console command: extractFeatures file_list.txt For matlab, % Function: % data = mexFeatures (root_folder, image_list, image_id, param, wght, ) % root foler: root path for these images, only need store the relative path % image_list: image file name list % image_id: assign an image id for each image in the list, will use it for matching purpose % parameters: [1] paramters for color feature extraction % [2-4]: color feature options: 1.HSV color histogram; 2.HSV 3D color histogram % 3.RGB Color moment 4.Lab color moment % for color histogram: number of bins for 2D/3D histogram, note as hb, sb, vb; % for color moment: number of x/y grid, param[3] is undefined in this case; % param[5]- size of Gabor filter mask, default ***; % param[6]- number of scale of Gabor filter, default 5; % param[7]- number of orientation of Gabor filter, default 8; % param[8]- control the sample rate for global representation, unused for Gabor moment, default 8; % param[9-12] GIST option % param[9]- number of GIST scale 4 % param[10]- filter mask size is 256 % param[11]- number of block to extract moment 4 % param[12]- fc for preprocessing, filter the image % wght: the weight to balance the contribution for each component, color, edge, Gabor, LBP and GIST, % the coefficient is manually assigned, roughly 1, 1, 1, 20, 7. The optimization of these coefficients is % knownn as "ensemble learning". % data: feature matrix with size of (total number images) x (feature length) % Here a simple example on the usage Folder = 'd:\data\images'; % image input folder Iname = dir([Folder '\*.jpg']); % set the parameters param = [3 3 3 0 *** 5 8 8 4 256 4 4]; wght = [1 1 1 20 7]; % Simple example, only extract two images J{1} = Iname(1).name; J{2} = Iname(2).name; % run extraction data = mexFeatures( Folder, J, 1:2, param, wght ); E: Normalization Note normalization is essential to obtaining the good results on a large collection. We suggest to use the following code to normalize the feature vector in C: void normalizeVector( double* v, int vSize) { double a, sqsum = 0.0; double mean = MeanVector( v, vSize); double* ptr = v; int i; for(i=0;i zim Hao Ma

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：