update

Author: niphy

ClassfyPictureTest.m

@@ -0,0 +1,15 @@
+function predicted_label = ClassfyPictureTest(testpath,ext,svmtestpath,n,kmeansfunc,model)
+%生成svmttext文件数据
+
+
+[telabel,tedata]=creatsvmfrompath(testpath,ext,svmtestpath,n,kmeansfunc);
+%进行预测
+%[predicted_label,accuracy,prob_estimates]=svmpredict(telabel,tedata,model);
+predicted_label = svmclassify(model,tedata);
+
+% matwrite('labels.txt',predicted_label);
+% matwrite('accuracy.txt',accuracy);
+% matwrite('prob_estimates.txt',prob_estimates);
+%下面的要自己统计（自己做了）
+
+end

ClassfyPictureTrain.m

@@ -0,0 +1,17 @@
+function model = ClassfyPictureTrain(trainpath,ext,svmtrainpath,n,kmeansfunc)
+%%通过trainpath内图像与kmeansfunc方式进行模型构建，获得model并保存
+%picturepath保存各个类别的图形的文件路径
+%kmeansfunc进行kmeans分类方式
+%%实现过程
+%生成svmtrain文件数据
+[trlabel,trdata]=creatsvmfrompath(trainpath,ext,svmtrainpath,n,kmeansfunc);
+%读取svmtrain数据
+%!python grid.py svmpath
+%得到c g 
+%0.03125 0.0078125 25.0 
+%预测模型
+model=svmtrain(trdata,trlabel);
+end
+
+
+

accuracy.txt

@@ -0,0 +1,3 @@
+1.000000	50.000000	
+1.000000	0.500000	
+1.000000	NaN	

creatsvmdata.m

@@ -0,0 +1,19 @@
+function creatsvmdata(idens,des,path)
+%%
+%idens为图片类别号
+%des为图片描述
+%path为保存路径
+%%
+%获得图片数目
+picsz=size(idens,1);
+picdsz=size(des,2);
+fid=fopen(path,'w');
+for i=1:picsz
+    fprintf(fid,'%d\t',idens(i));
+    for j=1:picdsz
+        fprintf(fid,'%d:%f\t',j,double(des(i,j)));
+    end
+    fprintf(fid,'\n');
+end
+fclose(fid);
+end

creatsvmfrompath.m

@@ -0,0 +1,21 @@
+function [palliden,palldes]=creatsvmfrompath(srcpaths,ext,savepath,n,kmeansfunc)
+%%由路径创建图片的svmdata的数据文件
+%srcpath保存各个类别的图形的文件路径
+%savepath保存svmdata的文件路径
+%kmeansfunc进行kmeans分类方式
+%%
+%获得分类数
+pclassnum=size(srcpaths,2);
+palliden=[];
+palldes=[];
+%进行各分类赋值
+for i=1:pclassnum
+    %获得一路径下所有文件特征
+    [idens,des]=getpathpicdes(srcpaths{1,i},ext,i,n,kmeansfunc);
+    %附加新值
+    palliden=[palliden;idens];
+    palldes=[palldes;des];
+end
+%保存训练数据
+creatsvmdata(palliden,palldes,savepath);
+end

fromkmeansresult.m

@@ -0,0 +1,22 @@
+function despic=fromkmeansresult(siftimg,n,IDX, C, sumd, D)
+%统计各聚类中心点的数目
+allcount(1,1:n)=0;
+potsz=size(IDX,1);
+for i=1:potsz
+    allcount(1,IDX(i))=allcount(1,IDX(i))+1;
+end
+maxcount=max(allcount);
+for i=1:n
+    if maxcount==allcount(i);
+        maxindex=i;
+    end
+end
+%统计maxnidexclus
+allclus=sum(C(maxindex,:));
+%获得sumd
+sumd=sumd';
+%获得各点距中心和
+allcenter=sum(D,1);
+%图片特征值返回
+despic=[allcount,allclus,sumd,allcenter];
+end

fromsiftresult.m

@@ -0,0 +1,33 @@
+function siftdes=fromsiftresult(siftpot,siftdes)
+potsz=size(siftpot,1);
+ad=[];
+for i=1:potsz
+    ang(1,1:potsz)=0;
+    dis(1,1:potsz)=0;
+    for j=1:potsz
+        x=siftpot(j,1)-siftpot(i,1);
+        y=siftpot(j,2)-siftpot(i,2);
+        dis(1,j)=sqrt(x*x+y*y);
+        ang(1,j)=atan2(y,x);
+    end
+    mang=mean(ang);
+    mdis=mean(dis);
+    diss(1,1:4)=0;
+    for j=1:potsz
+        if ang(1,j)>=mang
+            diss(1,1)=diss(1,1)+1;
+        end
+        if ang(1,j)<mang
+            diss(1,2)=diss(1,2)+1;
+        end
+        if dis(1,j)>=mdis
+            diss(1,3)=diss(1,3)+1;
+        end
+        if dis(1,j)<mdis
+            diss(1,4)=diss(1,4)+1;
+        end
+    end
+    ad=[ad;diss];
+end
+siftdes=[siftpot(:,3:4),ad];
+end

getallfiles.m

@@ -0,0 +1,10 @@
+function [filenames]=getallfiles(folderdir,ext)
+%%获得folderdir下的以ext为后缀的所有文件
+fileFolder=fullfile(folderdir);
+dirOutput=dir(fullfile(fileFolder,ext));
+filenames={dirOutput.name}';
+num=size(filenames,1);
+for i=1:num
+    filenames{i}=strcat(folderdir,filenames{i});
+end
+end

getpathpicdes.m

@@ -0,0 +1,18 @@
+function [cliden,cldes]=getpathpicdes(picturespath,ext,type,n,kmeanfunc)
+%%获得picturespath这一路径下所有图片的特征
+%picturespath路径
+%type分类号
+%kmeanfunc分类方式
+%%
+%获得所有图片路径
+filenames=getallfiles(picturespath,ext);
+%获得所有图片数目
+clnum=size(filenames,1);
+%获得特征Idens
+cliden=type*ones(clnum,1);%type是标签
+%构造图像特征
+cldes=[];
+for i=1:clnum
+    cldes=[cldes;picidening(filenames{i},n,kmeanfunc)];
+end
+end

labels.txt

@@ -0,0 +1,22 @@
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	
+1.000000	2.000000	

main.m

@@ -0,0 +1,2 @@
+trains = {'accordion','airplanes'};
+predicted_label = usingpro(trains,trains,'*.jpg',2,[]);

mat.txt

@@ -0,0 +1,5 @@
+1.000000	0.433938	2.000000	0.785353	3.000000	0.106040	4.000000	0.385658	5.000000	0.961285	
+1.000000	0.045593	2.000000	0.433492	3.000000	0.006074	4.000000	0.856946	5.000000	0.822906	
+1.000000	0.655473	2.000000	0.347821	3.000000	0.275042	4.000000	0.674825	5.000000	0.010499	
+1.000000	0.781042	2.000000	0.055389	3.000000	0.942771	4.000000	0.608097	5.000000	0.141617	
+1.000000	0.862900	2.000000	0.819441	3.000000	0.674259	4.000000	0.439426	5.000000	0.482810	

matwrite.m

@@ -0,0 +1,17 @@
+function matwrite(path,mat)
+%%
+%mat为Matrix
+%path为保存路径
+%%
+%获得图片数目
+picsz=size(mat,1);
+picdsz=size(mat,2);
+fid=fopen(path,'w');
+for i=1:picsz
+    for j=1:picdsz
+        fprintf(fid,'%f\t',j,double(mat(i,j)));
+    end
+    fprintf(fid,'\n');
+end
+fclose(fid);
+end

picidening.m

@@ -0,0 +1,33 @@
+function despic=picidening(picpath,n,kmeansfunc)
+%%获得pic的sift聚类特征
+%picpath为图片路径
+%n聚类数目
+%kmeansfunc聚类参数
+%%过程
+%获得sift特征点与特征点向量
+[siftimg,siftpot,siftdes]=sift(picpath);
+%由m，theta，与128维特征计算sift特征点相关性
+alldes=fromsiftresult(siftpot,siftdes);
+%进行基于特征点及特征的用kmeanfunc参数进行kmeans分类
+ifsz=size(kmeansfunc,2);
+if floor(ifsz/2)~=ifsz/2
+    error(message('kmeansfunc必须是一一对应的参数'));
+end
+if ifsz==0
+    [IDX,C,sumd,D]=kmeans(alldes,n);
+end
+if ifsz==2
+    [IDX,C,sumd,D]=kmeans(alldes,n,kmeansfunc{1,1},kmeansfunc{1,2});
+end
+if ifsz==4
+    [IDX,C,sumd,D]=kmeans(alldes,n,kmeansfunc{1,1},kmeansfunc{1,2},kmeansfunc{1,3},kmeansfunc{1,4});
+end
+if ifsz==6
+    [IDX,C,sumd,D]=kmeans(alldes,n,kmeansfunc{1,1},kmeansfunc{1,2},kmeansfunc{1,3},kmeansfunc{1,4},kmeansfunc{1,5},kmeansfunc{1,6});
+end
+if ifsz==8
+    [IDX,C,sumd,D]=kmeans(alldes,n,kmeansfunc{1,1},kmeansfunc{1,2},kmeansfunc{1,3},kmeansfunc{1,4},kmeansfunc{1,5},kmeansfunc{1,6},kmeansfunc{1,7},kmeansfunc{1,8});
+end
+%据分类结果计算特征量
+despic=fromkmeansresult(siftimg,n,IDX, C, sumd, D);
+end

prob_estimates.txt

@@ -0,0 +1,22 @@
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	
+1.000000	0.000000	

sift.m

@@ -0,0 +1,88 @@
+% [image, descriptors, locs] = sift(imageFile)
+%
+% This function reads an image and returns its SIFT keypoints.
+%   Input parameters:
+%     imageFile: the file name for the image.
+%
+%   Returned:
+%     image: the image array in double format
+%     descriptors: a K-by-128 matrix, where each row gives an invariant
+%         descriptor for one of the K keypoints.  The descriptor is a vector
+%         of 128 values normalized to unit length.
+%     locs: K-by-4 matrix, in which each row has the 4 values for a
+%         keypoint location (row, column, scale, orientation).  The 
+%         orientation is in the range [-PI, PI] radians.
+%
+% Credits: Thanks for initial version of this program to D. Alvaro and 
+%          J.J. Guerrero, Universidad de Zaragoza (modified by D. Lowe)
+
+function [image, descriptors, locs] = sift(imageFile)
+
+% Load image
+
+image = imread(imageFile);
+
+% If you have the Image Processing Toolbox, you can uncomment the following
+%   lines to allow input of color images, which will be converted to grayscale.
+if ndims(image) == 3
+    image = rgb2gray(image);
+end
+
+[rows, cols] = size(image); 
+
+% Convert into PGM imagefile, readable by "keypoints" executable
+f = fopen('tmp.pgm', 'w');
+if f == -1
+    error('Could not create file tmp.pgm.');
+end
+fprintf(f, 'P5\n%d\n%d\n255\n', cols, rows);
+fwrite(f, image, 'uint8');
+fclose(f);
+
+% Call keypoints executable
+if isunix
+    command = '!./sift ';
+else
+    command = '!siftWin32.exe ';
+end
+command = [command ' <tmp.pgm >tmp.key'];
+eval(command);
+
+% Open tmp.key and check its header
+g = fopen('tmp.key', 'r');
+if g == -1
+    error('Could not open file tmp.key.');
+end
+[header, count] = fscanf(g, '%d %d', [1 2]);
+if count ~= 2
+    error('Invalid keypoint file beginning.');
+end
+num = header(1);
+len = header(2);
+if len ~= 128
+    error('Keypoint descriptor length invalid (should be 128).');
+end
+
+% Creates the two output matrices (use known size for efficiency)
+locs = double(zeros(num, 4));
+descriptors = double(zeros(num, 128));
+
+% Parse tmp.key
+for i = 1:num
+    [vector, count] = fscanf(g, '%f %f %f %f', [1 4]); %row col scale ori
+    if count ~= 4
+        error('Invalid keypoint file format');
+    end
+    locs(i, :) = vector(1, :);
+    
+    [descrip, count] = fscanf(g, '%d', [1 len]);
+    if (count ~= 128)
+        error('Invalid keypoint file value.');
+    end
+    % Normalize each input vector to unit length
+    descrip = descrip / sqrt(sum(descrip.^2));
+    descriptors(i, :) = descrip(1, :);
+end
+fclose(g);
+
+