diff --git a/face_video.mp4 b/face_video.mp4
new file mode 100644
index 0000000..d4f0581
Binary files /dev/null and b/face_video.mp4 differ
diff --git a/frame_extraction.py b/frame_extraction.py
new file mode 100644
index 0000000..956a452
--- /dev/null
+++ b/frame_extraction.py
@@ -0,0 +1,19 @@
+############
+# frame extraction
+# 
+# Task:
+#	-frame extration from mp4 video
+#
+# Inputs:
+#	-mp4 file
+# 	-fps
+#	-
+#	-
+#
+# Output:
+#	-png of video
+#
+# Author: Thomas Périn
+# Date: 17/02/2023
+#
+############
diff --git a/heartRateEstimation.m b/heartRateEstimation.m
new file mode 100644
index 0000000..9d55adb
--- /dev/null
+++ b/heartRateEstimation.m
@@ -0,0 +1,80 @@
+%hr = heartRateEstimation(imgDirectory, fps, windowDuration, windowShift)
+%%%%%%%%%%%%%%%%%
+% hr = heartRateEstimation(imgDirectory, fps, windowDuration, windowShift)
+% 
+% Task:
+%
+% Inputs:
+%	-imgDirectory
+% 	-fps
+%	-windowDuration
+%	-windowShift
+%
+% Output:
+%	-hr:
+%
+% Author: Guillaume Gibert
+% Date: 06/02/2023
+%
+% Note: images were extracted with ffmpeg -i "video.mkv" ../img/gg_%04d.png
+%%%%%%%%%%%%%%%%%
+
+%temp var
+imgDirectory = 'img';
+fps = 60;
+windowDuration = 30;
+windowShift = 1;
+roi = [225 750 825 1100];
+
+% global 
+redChannel = [];
+greenChannel = [];
+blueChannel = [];
+
+%list the images
+listImg = dir([imgDirectory '/*.png']);
+
+for l_img=1:windowDuration*fps
+	% load the current image
+	image_original = imread([imgDirectory '/' listImg(l_img).name]);
+	%image(image_original);
+	
+	% crop the current image around the face
+	image_face = image_original(roi(1):roi(2), roi(3):roi(4), :);
+	%image(image_face);
+	
+	% spatial average for r, g, b channels
+	r_mean = mean(mean(image_face(:,:,1)));
+	g_mean = mean(mean(image_face(:,:,2)));
+	b_mean = mean(mean(image_face(:,:,3)));
+	
+	% store the current "average" pixel in global arrays
+	redChannel = [redChannel r_mean];
+	greenChannel = [greenChannel g_mean];
+	blueChannel = [blueChannel b_mean];
+end
+
+% estimate temporal average and standard deviation
+greenChannel_avg = mean(greenChannel);
+greenChannel_std = std(greenChannel);
+
+% normalize your data 
+% greenChannel_normalized(i) = (greenChannel(i) - greenChannel_avg)/greenChannel_std
+
+% fft(greenChannel_normalized)
+
+% power spectrum (https://www.mathworks.com/help/matlab/ref/fft.html)
+
+% find the peak in the range ([0.75 4] Hz)
+% max -> value, index
+
+% convert the index from Hz to bpm
+
+% determine heart rate
+
+
+
+
+
+
+