how to plot confusionmat for this code?
1 Ansicht (letzte 30 Tage)
Ältere Kommentare anzeigen
close all clear all clc delete('Frames\*.jpg'); [filename pathname] = uigetfile({'*.avi'},'Select A Video File'); I = VideoReader([pathname,filename]); implay([pathname,filename]); pause(3); nFrames = I.numberofFrames; vidHeight = I.Height; vidWidth = I.Width; mov(1:nFrames) = ... struct('cdata', zeros(vidHeight, vidWidth, 3, 'uint8'),... 'colormap', []); WantedFrames = 50; for k = 1:WantedFrames mov(k).cdata = read( I, k); mov(k).cdata = imresize(mov(k).cdata,[256,256]); imwrite(mov(k).cdata,['Frames\',num2str(k),'.jpg']); end
for I = 1:WantedFrames im=imread(['Frames\',num2str(I),'.jpg']); figure(1),subplot(5,10,I),imshow(im); end clc for i=1:WantedFrames disp(['Processing frame no.',num2str(i)]); img=imread(['Frames\',num2str(i),'.jpg']); f1=il_rgb2gray(double(img)); [ysize,xsize]=size(f1); nptsmax=40; kparam=0.04; pointtype=1; sxl2=4; sxi2=2*sxl2; % detect points [posinit,valinit]=STIP(f1,kparam,sxl2,sxi2,pointtype,nptsmax); Test_Feat(i,1:40)=valinit; %imshow(f1,[]), hold on % axis off; % showellipticfeatures(posinit,[1 1 0]); % title('Feature Points','fontsize',12,'fontname','Times New Roman','color','Black') end
% Use KNN To classify the videos load('TrainFeature.mat') X = meas; Y = New_Label; Z = Test_Feat; % Now Classify
%ens = fitensemble(X,Y,'Subspace',300,'KNN'); %class = predict(ens,Z(1,:)) md1 = ClassificationKNN.fit(X,Y); Type = predict(md1,Z); if (Type == 1) disp('Boxing'); helpdlg(' Boxing '); elseif (Type == 2) disp('Hand Clapping'); helpdlg('Hand Clapping'); elseif (Type == 3) disp('Hand Waving'); helpdlg('Hand Waving'); elseif (Type == 4) disp('Jogging'); helpdlg('Jogging'); elseif (Type == 5) disp('Running'); helpdlg('Running'); elseif (Type == 6) disp('Walking'); helpdlg('Walking'); elseif (Type == 7) disp('Cycling'); helpdlg('Cycling'); elseif (Type == 8) disp('Surfing'); helpdlg('Surfing'); end
0 Kommentare
Antworten (1)
Yuvaraj Venkataswamy
am 18 Jun. 2018
if true
plotconfusion(actual_labels,Predicted_labels)
end
In this, Predicted_labels are which you have classified through KNN and actual_labels are the true labels.
0 Kommentare
Siehe auch
Kategorien
Mehr zu Text Data Preparation finden Sie in Help Center und File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!