how to save feature vector of SURF for each image and train, test it for classification to svm?

output for image 1: 322*64

output for image 2: 593*64

and so on the value*64 changes for each image. can anyone please help me how to save so many features and then use it for training.

Answer 1

[KALYAN ACHARJYA]

"output for image 1: 322*64

output for image 2: 593*64

and so on the value*64 changes for each image. can anyone please help me how to save so many features and then use it for training"

Use cell array, I am answering the question to save different sizes/types results data

I am assuming you are calling Images

result=cells(1,total_images);
%................^Total result images 
for i=1:total_images
%image read
result{i}= output result;
end

[madhan ravi]

Preallocate the cell array properly.

[chetna kaushal]

Sir how to feed them for testing.

[KALYAN ACHARJYA]

Comment to @Madhan, As I have no idea the number of iteration?? Any Clue?

Thanks

[KALYAN ACHARJYA]

@Chetna, that part I have no idea, I never worked on it, but certainly it seems easy steps. it would be better open new question, and experts will answer on it.

[madhan ravi]

for i = 1 : total_images

[KALYAN ACHARJYA]

Ok @Madhan Thanks

[chetna kaushal]

training = 'image';

filenames = dir(fullfile(training_female, '*.png'));

total_images = numel(filenames);

featureMatrix = [];

for n = 1:total_images

full_name= fullfile(training, filenames(n).name);

training_images = imread(full_name);

training_images=rgb2gray(training_images);

Fpoints = detectSURFFeatures(training_images);

[features, validPoints] = extractFeatures(training_images,Fpoints);

featurematrix{n} = [featureMatrix; features];

suppose I have 8 images.

The featuematrix has 8 cell struct. with different values in a 64 column for each image.

how do i now label them and then use them for classification. how do i call them for testing.