hello, i wanna do iris recognition by compare an iris image template on axes1 with a folder that filled with iris image templates, the iris image template from the folder with lowest hamming distance value will be the output and displayed on axes2.

here is the codes that i have so far:

function buttonpencocokan_Callback(hObject, eventdata, handles)
tic;
gettraining=get(handles.txtketlatih, 'String');
datapath = (gettraining);
testpath = ('F:\Skripsi\Citra Uji');
s = get(handles.txtcitrauji, 'String');
file = char(s);
[pathstr,name,ext] = fileparts(file);
TestImage = strcat([pathstr],'\',[name,ext]);

recog_img = hammingdist(datapath,TestImage);
selected_img = strcat(datapath,'\',recog_img);
select_img = imread(selected_img);
test_img = imread(TestImage);

axes(handles.axes2);
imagesc(select_img);
set(handles.axes2, 'Visible', 'Off');
result = strcat(recog_img);
set(handles.txtnmlatih,'String',result);

conn=database('database','root','','com.mysql.jdbc.Driver','jdbc:mysql://localhost:3306/');
S = get(handles.txtnmlatih, 'string');
N = cell2mat(regexp(S, '\d+', 'match')); 
set(handles.txtid,'String',[N]);
idx = get(handles.txtid, 'string');
idxi = str2num(idx);
idx = num2str( ceil( idxi / 1) );
set(handles.txtid,'String',[idx]);
sqlquery = ['select * from irismata where id =' '''' idx ''''];
biodata = exec(conn,sqlquery);
biodata = fetch(biodata);
biodata.Data;
nama=biodata.Data{1,2};
set(handles.txtnmuji,'String',nama);
lokasi=biodata.Data{1,3};
set(handles.txtlokasiuji,'String',lokasi);
t=toc;
set(handles.txtwaktu,'string',num2str(t))

ambil = get(handles.txtnmuji, 'string');
set(handles.txtketuji, 'String', ambil);
ambil2 = get(handles.txtnama, 'string');
set(handles.txtketlatih, 'String', ambil2);

close(conn);
set(handles.HasilEktraksi,'Enable','on');

[Walter Roberson]

What is your question?

[Glenn]

hahah sorry sir i forgot, it says: Error using dir Invalid path. The path must not contain a null character.

Error in imread (line 346) if ~isempty(dir(filename))

Answer 1

[Walter Roberson]

Same basic problem as https://www.mathworks.com/matlabcentral/answers/341843-load-files-from-relative-path

You have

recog_img = hammingdistance(datapath,TestImage);

Mathworks does not provide a hammingdistance function so I do not know exactly what the function returns, but it is pretty likely that it is a numeric value, possibly an integer. In the situation of an exact match, we can guess that the value it would return would be 0, by definition of hamming distance.

You then have

selected_img = strcat(datapath,'\',recog_img);

Now in the case where recog_img is indeed numeric, then this would be equivalent to

selected_img = strcat(datapath, '\', char(recog_img) );

and in particular when recog_img is numeric 0, this would be

selected_img = strcat(datapath, '\', char(0) );

char(0), the very first character in the ASCII (or ISO, or Unicode) character sets, is formally known as NUL, or less formally as the null character -- much the same way that char(13) is formally known as CR, or less formally as carriage return. And the thing about char(0), NUL, is that it is not permitted in a file name in MS Windows -- not in the shells, not in the FAT file system, nor in NTFS file system. So when you accidentally put one into a file name, the program complains.

char(0) is not the same thing as '0' . Each character has an associated position in the character code tables, and it is the positions that are computed with rather than the graphic representation of the character. The graphic representation of the digit for zero, '0', is at position #48, for various reasons.

If you want numeric 0 to become '0' for the purpose of selected_img you need to convert the numeric to string, such as

    selected_img = fullfile(datapath, num2str(recog_img) );

[Glenn]

hammingdist actually a .m file that i add into my project folder. its for counting the hamming distance value between template on axes1 and the templates in the "Citra Latih" folder, the template from the folder that has lowest hamming distance value compared to the template on axes1 will be the output and displayed on axes2. hamming distance value determines the value of the difference between compared templates, the lowest possible value is 0 means its the same template. the highest value could be any number (wont be more than 0.999 i guess, so it will be a double or float instead of an integer). by the way, i did what u suggested me to do:

function buttonpencocokan_Callback(hObject, eventdata, handles) 
tic; 
gettraining=get(handles.txtlokasilatih, 'String'); 
datapath = (gettraining);

TestImage=getappdata(handles.axes1, 'templateVVV'); 
recog_img = hammingdist(datapath,TestImage);  
selected_img = strcat(datapath, '\', char(0) ); 
select_img = imread(selected_img); 
test_img = imread(TestImage); 

axes(handles.axes2); 
imagesc(select_img); 
set(handles.axes2, 'Visible', 'Off'); 
result = strcat(recog_img); 
set(handles.txtnmlatih,'String',result); 

conn=database('database','root','','com.mysql.jdbc.Driver','jdbc:mysql://localhost:3306/'); 
S = get(handles.txtnmlatih, 'string'); 
N = cell2mat(regexp(S, '\d+', 'match'));  
set(handles.txtid,'String',[N]); 
idx = get(handles.txtid, 'string'); 
idxi = str2num(idx); 
idx = num2str( ceil( idxi / 1) ); 
set(handles.txtid,'String',[idx]); 
sqlquery = ['select * from irismata where id =' '''' idx '''']; 
biodata = exec(conn,sqlquery); 
biodata = fetch(biodata); 
biodata.Data; 
nama=biodata.Data{1,2}; 
set(handles.txtnmuji,'String',nama); 
lokasi=biodata.Data{1,3}; 
set(handles.txtlokasilatih,'String',lokasi); 
t=toc; 
set(handles.txtwaktu,'string',num2str(t)) 

ambil = get(handles.txtnmuji, 'string'); 
set(handles.txtketuji, 'String', ambil); 
ambil2 = get(handles.txtnama, 'string'); 
set(handles.txtketlatih, 'String', ambil2); 

close(conn); 
set(handles.HasilEktraksi,'Enable','on');

and it says:

Error using imread (line 347)
Cannot open file "F:\Skripsi\Citra Latih\" for reading. You might not have read permission.

sorry for my terrible english sir. :D

[Walter Roberson]

Change

selected_img = strcat(datapath, '\', char(0) );

to

selected_img = fullfile(datapath, sprintf('%04d.png', floor(1000*recog_img) ) );

This assumes that you want to construct the name of the selected image by taking the hamming distance and multiplying by 1000 and taking the integer part, and using the 4 digit number from that with the file extension .png . This would, for example, construct the file name 0317.png for hamming distances 0.03179218 . This filename construction is a guess at your needs. What do the file names look like?

[Glenn]

did you mean the file name in the folder? could be any character like sample7_1.bmp, ... , sample7_4.bmp. when i changed

 selected_img = strcat(datapath, '\', char(0) );

to

      selected_img = fullfile(datapath, sprintf('%04d.bmp', floor(1000*recog_img) ) );

it says:

Error using imread (line 349) File "F:\Skripsi\Citra Latih\0500.bmp" does not exist.

ah so the hamming distance value from that comparation is 0.500? can the value displayed on a text field instead? the text field name is txtnilaikemiripan.

[Image Analyst]

Either use dir() to get a list of files that actually are guaranteed to exist, or use the exist() function to skip the file if it doesn't exist

if ~exist(selected_img, 'file')
    continue;  % Skip to bottom of loop and continue with the next image.
end

[Walter Roberson]

Okay, so when the hamming distances comes out as 0.500, then which file name should that correspond to? You want to use the hamming distances as part of the file name, but how do you convert between hamming distance and (for example) sample7_4.bmp ? Is the '_4' intended to indicate hamming distance 4 and then we just need to figure out how to convert between the relative hamming distance 0.5 and the absolute hamming distance 4? (If that is the case, multiple the relative hamming distance by the length of the input to get the absolute hamming distance.)

A file name like sample7_4.bmp has two numeric components. Which one encodes the hamming distance? And how is the other one to be known? Is there possibly also a sample6_4.bmp ? If so then how do you know if you want sample7_4.bmp or sample6_4.bmp ?

[Glenn]

no, i dont want the hamming distance value as the filename sir. the file in the folder already has a name. i want the value to be displayed on the text field as an information later. the steps are: there is axes1 that contains an iris image template. then i will push recognize button (the codes i wrote in this question), the process in this button should be comparing templates on axes1 to a folder that contains image templates(folder Citra Latih). there are alot of template in the folder. the template on axes1 will be compared one by one with the template from the folder by using hamming distance. the lowest value from the one by one comparing process will be the output template and will be displayed on axes2. there will be informations on the text fields such as the time needed to recognize the iris template, the output template file name, the hamming distance value(must be the lowest value from the one by one comparation), etc. ah is it needed to store all the one by one comparation result before choosing the lowest value amongs them? i have no idea how to do it sir.

[Walter Roberson]

Your code had

recog_img = hammingdist(datapath,TestImage);
selected_img = strcat(datapath,'\',recog_img);

That tries to use the hamming distance value as part of the name of the image file to select.

[Glenn]

no the name will be displayed on a textfield.

result = strcat(recog_img); set(handles.txtnmlatih,'String',result);

[Glenn]

i found the solution based on your answer sir, thanks alot for your help. God bless you.