calibration in image analysis

find (BW==1);

[m,n]= find (BW==1);

pos = [m';n'];

D = zeros(1,448)

for i = 1:length(m)

i

D(i,:) = [calibrated_data(m(i),n(i),:)]

end

i have used the above code for calibration but the dataset has 398656 elements and it is taking so long to calibrate like more than 1hr. is there any faster way for calibration??

[Benjamin Thompson]

Can you provide more information about what you are calibrating, and maybe a small sample image? What is the problem?

[Chaitanya kumar reddy]

clc

clear all

close all

dataset_name = ['20220118_test_Normalbreast_061722_1_left_2022-06-17_19-28-26'];

raw_data_headfile = strcat('D:\HSI_data\', dataset_name, '\capture\', dataset_name, '.hdr');

raw_data_datafile = strcat('D:\HSI_data\', dataset_name, '\capture\', dataset_name, '.raw');

dark_ref_headfile = strcat('D:\HSI_data\', dataset_name, '\capture\DARKREF_', dataset_name, '.hdr');

dark_ref_datafile = strcat('D:\HSI_data\', dataset_name, '\capture\DARKREF_', dataset_name, '.raw');

white_ref_headfile = strcat('D:\HSI_data\', dataset_name, '\capture\WHITEREF_', dataset_name, '.hdr');

white_ref_datafile = strcat('D:\HSI_data\', dataset_name, '\capture\WHITEREF_', dataset_name, '.raw');

raw_info = enviinfo(raw_data_headfile);

raw_data = multibandread(raw_data_datafile,[raw_info.Height, raw_info.Width, raw_info.Bands],...

raw_info.DataType, raw_info.HeaderOffset, raw_info.Interleave, raw_info.ByteOrder);

dark_info = enviinfo(dark_ref_headfile);

dark_data = multibandread(dark_ref_datafile, [dark_info.Height, dark_info.Width, dark_info.Bands],...

dark_info.DataType, dark_info.HeaderOffset, dark_info.Interleave, dark_info.ByteOrder);

dark_data_avg = mean(dark_data, 1);

white_info = enviinfo(white_ref_headfile);

white_data = multibandread(white_ref_datafile, [white_info.Height, white_info.Width, white_info.Bands],...

white_info.DataType, white_info.HeaderOffset, white_info.Interleave, white_info.ByteOrder);

white_data_avg = mean(white_data, 1);

calibrated_data = raw_data;

for i = 1:size(raw_data, 1)

calibrated_data(i, :, :) = (calibrated_data(i, :, :) - dark_data_avg)./ (white_data_avg - dark_data_avg);

end

rice_dot = calibrated_data(238, 460, :);

background_dot = calibrated_data(323, 658, :);

plot(raw_info.Wavelength, rice_dot(:))

hold on

plot(raw_info.Wavelength, background_dot(:))

legend('grape', 'background')

A = imread('D:\HSI_data\20220118_test_Normalbreast_061722_1_left_2022-06-17_19-28-26\20220118_test_Normalbreast_061722_1_left_2022-06-17_19-28-26.png');

imshow(A)

B = rgb2gray(A);

level = graythresh(B);

BW = im2bw(B,level);

find (BW==1);

[m,n]= find (BW==1);

pos = [m';n'];

D = zeros(1,448)

%D = cell([1 2]);

for i = 1:length(m)

D(i,:) = [calibrated_data(m(i),n(i),:)];

end

% E = mean(D);

% plot(raw_info.Wavelength,E)

% figure

% imshow(BW);

%

%

% % D = calibrated_data(70,275,:);

% % E = reshape(D,[1,448]);

Answer 1

[Udit06]

Hi Chaitanya,

I understand that the code that you have written for calibration is taking a long time to run. You can utilize the following features supported by MATLAB to make your code run faster.

1) Vectorization: Instead of using a loop to iterate over each element, try to use vectorized operations whenever possible. MATLAB is optimized for array operations, and using vectorization can significantly speed up the computation.

2) Parallel Computing: Consider using parallel computing techniques like MATLAB's "parfor" to distribute the workload across multiple workers.

You can refer to the below MathWorks documentations to understand more about vectorization and “parfor” respectively.

1) https://www.mathworks.com/help/matlab/matlab_prog/vectorization.html searchHighlight=vectorization&s_tid=srchtitle_support_results_1_vectorization

2) https://www.mathworks.com/help/parallel-computing/parfor.html

I hope this helps.