This bachelor thesis gives an overview of human skin, different skin spots and the aging process. Within a practical work it was looked after how these properties can be adopted in the computer-generated visualization.
For this purpose, skin was analysed without skin spots and other alterations. Three different methods were used to apply the previously acquired knowledge in practice and create prototypes. Those were then tested on how well they would work in a system.
In a further step an UI-Design was developed as a proof of concept to demonstrate the functionality of the system.
The aim was to find the best method in regard to the result and the needed effort, to be used in a system like this.
A big problem was finding suitable literature. While searching for artistic aspects of skin, every found book described the biologic aspect on how skin is built and how the process of aging goes on, rather than what the signs of aging are or how changes look like. With the information from a few books in combination with some cosmetic surgery websites, it was possible to gather the necessary information which was required for the pending work.
One of the most important insights was to find out, that the human skin is divided into two different patterns. While the first has a rhombic pattern on 96% of the skin surface, the second pattern runs across the remaining 4% of the body in lines. (You can see this on the two pictures below. The first picture shows the rhombic pattern, the second shows the line pattern, which can be found on the palm of the hands and feet).
Another part of the theoretical work was to look after different skin conditions and types and to find out other minor peculiarities of the skin. These include various skin spots which represent the human appearance as we know it. This work focused on freckles, liver spots and aging spots.
The last big part of the work was to find out more about the aging process and the different forms coming with it. While most of the work concentrated on the aging of the face, attention was also paid to the difference in skin aging with different ethnic groups.
A topic that was left out in this thesis was body hair.
For the prototype which would be created with the help of photogrammetry, a couple of test pictures were taken. These pictures showed quickly that the arm did not remain calm resulting in blurry pictures. Also, clues were no longer at the same place.
Another method had to be found. To develop sharp pictures the idea of body contouring came up. With a special prepared silicone, which is made for body contouring, a part of the arm could be moulded. On this silicone fine details were recognizable. After taking test pictures again bubbles where visible in the silicone, making these pictures useless. The next step was to turn the silicone into a gypsum imprint. Although the gypsum developed some bubbles, the pattern was very well recognizable in some places and therefore useful for further work.
After the gypsum imprint was a success Autodesk Photo ReCap could be used. After all data was generated in ReCap, it could be further processed in Maya. From this data a normal and a displacement map was generated. Both were placed on a 1x1 cm plane to create the same render for every prototype. But the resulting render with this pattern looked to washed out. Even though young skin was taken for the imprint, it looked more like old skin.
Another method of creating a prototype was using the software “ShaderMaps 4”. This method was very efficient, and the resulting prototype showed every single fine detail.
Since there were already pictures, one prototype was made with a grayscale picture. For this purpose, one of these pictures was imported into Mudbox. The pattern was transferred to a plane and then exported as a displacement-map. The result looked like the skin pattern, but through the image there were still light information remaining inside the pattern, leading to shadows, where no shadows should be.
The fourth method was to model the pattern procedural. Therefore, different noises were used and blended together. This method worked well, but the resulting pattern does not quite resemble the pattern of skin. To achieve this, it needs to be redone or even an own noise needs to be programmed.
You can see some of the results here:
First, it should be examined whether the skin pattern is the same for every skin type. After that the pattern needs to be analysed for its appearance under the different skin conditions on the face. These different skin conditions also must be considered more closely to digitalise them.
The second skin pattern, which is found only on 4% of the body, needs to be digitalised too. It would be conceivable to program separate noises for each of these patterns. The final result should look more like real skin.
Another big aspect is the aging process. For this purpose, the skin of younger and older humans need to be analysed. For this step the method of using silicone and gypsum could be used again. Furthermore, a way must be found how crinkles can be placed on the digital body and face.
Rendered with Renderman.
You want more information about this project? Just send me a mail!