diff --git a/octave/PPG.m b/octave/PPG.m
new file mode 100644
index 0000000..b241e67
--- /dev/null
+++ b/octave/PPG.m
@@ -0,0 +1,35 @@
+% First we have to extract the values of red(R), green(G) and blue(B) from the txt file.
+fileID = fopen("listRGB.txt"); % Opening the txt file
+M = textscan(fileID, "[[%f, %f, %f]]"); % Extracting the values
+fclose(fileID); % Closing the file
+R = M{1};
+G = M{2};
+B = M{3};
+
+% Then we normalize the different RGB vectors but substracting the mean value and divide by the standard deviation.
+Normal_R = (R - mean(R)) ./ std(R);
+Normal_G = (G - mean(G)) ./ std(G);
+Normal_B = (B - mean(B)) ./ std(B);
+
+% We now use the fastICA functions in order to obtain 3 channels of 450 samples each.
+[icasig, A, W] = fastica([Normal_R'; Normal_G'; Normal_B']);
+
+% Here we compute the magnitude of the fast Fourrier's transform.
+PPG_spectrum_R = abs(fft(Normal_R));
+PPG_spectrum_G = abs(fft(Normal_G));
+PPG_spectrum_B = abs(fft(Normal_B));
+
+% This line of code is for computing the frequency of each samples.
+f = [0:450]./450*15;
+
+% Then we find the maximum of each channels and by knowing the index of it, we find the corresponding frequency.
+[value_R, idx_R] = max(PPG_spectrum_R);
+heart_rate_freq_R = f(idx_R);
+[value_G, idx_G] = max(PPG_spectrum_G);
+heart_rate_freq_G = f(idx_G);
+[value_B, idx_B] = max(PPG_spectrum_B);
+heart_rate_freq_B = f(idx_B);
+
+% Finally, we display the BPM by multiplying the frequency by 60 and find the maximum of each channels.
+disp(max([heart_rate_freq_R*60,heart_rate_freq_G*60,heart_rate_freq_B*60]))
+