Introduction

This program is an implementation of the paper "Individualized 3D Face Model Reconstruction using Two Orthogonal Face Images" by Weon et. al. by Hansjoerg Hofer (1026632) and Sebastian Metzler (0927550).

Execution

The program 3DSelfie.exe has to be called with the parameter "-force-opengl" in order to run correctly.
Alternatively you can use the attached shortcut 3DSelfie_StartMe.bat.

Note: A webcam is required!

The following is a step by step guide to generate your personal 3D Face Model:

Step 1: Front foto and face contour

As soon as the programm starts you can see the webcam image of yourself in the window. For the best results you should place yourself in front of a white background. Now try to alter you position so that your eyes and mouth fit the pink rectangles. The 6 sliders on the bottom are used to modify the YCbCr thresholds. The first two sliders set the minimal and maximal value for the brightness (Y). The third and fourth slider modify the range of the blue-yellow difference (Cb). End the last two sliders are responsible for the red-green difference (Cr)

In the first stage you have to change these sliders so that the background around your head is segmented. You can see that a part of the image is segmented when it turns darker. Be careful, there should not be any segmentations in your face, in this stage! When you have segmented your face contour correctly and your eyes and mouth fit the pink rectangles, press the button on the top.

Step 2: Segmentation of the Eyes

Now your foto is saved and you can move freely. In this step you have to segment your eyes. Change the sliders so that your eyes turn darker. Areas outside the pink rectangles around the eyes are ignored, so you can focus on the eyes. If you are satisfied with the segmentation, press the button again.

Step 3: Segmentation of the Eyebrows

This step focuses on the eyebrows. Segment the eyebrows and click the button.

Step 4: Segmentation of the Mouth

This step focuses on the mouth. Try to fill the whole lips area with a continuous area without holes. Segment the mouth and click the button.

Step 5: Side foto

Now the segmentation is done and you have to take another foto, this time turn your head to the left. The head should be completely visible. To take the foto, click the button.

Step 6: Binary image

Now you can lean back and watch the program do its work. This may take some time. The picture you see is the segmented binary image of your facial features. When the program is done, you can see the matched control points from the snake algorithm.

Step 7: View result

With the Z-key it is possible to switch to the generated face model. You can rotate the model by clicking and draggin.

Enjoy your 3D Selfie!