How can I detect the SURF features of 400 images? (I want the code)

Hello, can you please help me detect the SURF features of 400 images and stock them in a matrix [?? 400] or [400 ??]?

My code is

clear all, clc, close all;
a=dir('C:\Users\toufik\Desktop\database\ORL\*.jpg');
n= length(a);
for i=1:n
filename=strcat('C:\Users\toufik\Desktop\database\ORL\',a(i).name);
    I=imread(filename);
    points = detectSURFFeatures(I);
    [features, valid_points] = extractFeatures(I, points);
end

[Walter Roberson]

all_features(:, i) = features ;

[lebed toufik]

@Walter Roberson Thank you a lot for answering my question, but I have a problem.

"Subscripted assignment dimension mismatch."

[Walter Roberson]

all_features(:, :, i) = features ;

[lebed toufik]

@Walter Roberson same problem.

[Walter Roberson]

a = dir('C:\Users\toufik\Desktop\database\ORL\*.jpg');
n = length(a);
for i=1:n
    filename = strcat('C:\Users\toufik\Desktop\database\ORL\',a(i).name);
    I = imread(filename);
    points = detectSURFFeatures(I);
    [features{i}, valid_points{i}] = extractFeatures(I, points);
end

Note that each of the features will be a 2D array; you could reshape that to a row or column if needed.

Note that the number of points detected will depend upon the content of the image, so the features array may be a different size for each input file, even if the input files are all the same size (which is something your code does not promise.)

Answer 1

[yanqi liu]

yes，sir，may be you want use surf to classify orl databse，or search image by surf points，so we can use bagOfFeatures to get it，such as

a=dir('C:\Users\toufik\Desktop\database\ORL\*.jpg');
ab = imageSet('C:\Users\toufik\Desktop\database\ORL','recursive');
bag = bagOfFeatures(ab,'Verbose',false);
% get surf feature vector
img = imread(strcat('C:\Users\toufik\Desktop\database\ORL\',a(1).name););
featureVector = encode(bag,img);

[lebed toufik]

@yanqi liu Thank you a lot for your response, but I want to create a matrix in which each image has its own features for example [400 ??] or [?? 400].

[yanqi liu]

yes，sir，may be different image have different surf points number，so the description feature vector may be not the same size，so we can use vision code dictionary to make them same size or choose same number surf points to get same size

[lebed toufik]

@yanqi liu thank you a lot sir.