Skin Microstructure Deformation with Displacement Map Convolution
SIGGRAPH 2015 Computer Animation Festival    SIGGRAPH 2015 Technical Papers
Koki Nagano    Graham Fyffe    Oleg Alexander   
Jernej Barbic1    Hao Li1    Abhijeet Ghosh2    Paul Debevec
USC Institute for Creative Technologies    University of Southern California1    Imperial College London2


We present a technique for synthesizing the effects of skin microstructure deformation by anisotropically convolving a highresolution displacement map to match normal distribution changes in measured skin samples. We use a 10-micron resolution scanning technique to measure several in vivo skin samples as they are stretched and compressed in different directions, quantifying how stretching smooths the skin and compression makes it rougher. We tabulate the resulting surface normal distributions, and show that convolving a neutral skin microstructure displacement map with blurring and sharpening filters can mimic normal distribution changes and microstructure deformations. We implement the spatially-varying displacement map filtering on the GPU to interactively render the effects of dynamic microgeometry on animated faces obtained from high-resolution facial scans.

Figure 1: (left) Microstructure acquistion in a polarized LED sphere with macro camera and articulated skin deformer. (right) A rendered facial expression with (a) mesostructure only (b) static microstructure from a neutral expression (c) dynamic microstructure from convolving the neutral microstructure according to local surface strain compared to a (d) reference photograph of the similar expression. The insets show detail from the lower-left area.

Simulating the appearance of human skin is important for rendering realistic digital human characters for simulation, education, and entertainment applications. Skin exhibits great variation in color, surface roughness, and translucency over different parts of the body, between different individuals, and when it's transformed by articulation and deformation. But as variable as skin can be, human perception is remarkably attuned to the subtleties of skin appearance, as attested to by the vast array of makeup products designed to enhance and embellish it.

Advances in measuring and simulating the scattering of light beneath the surface of the skin have made it possible to render convincingly realistic human characters whose skin appear to be fleshy and organic. Today's high-resolution facial scanning techniques (e.g. record facial geometry, surface coloration, and surface mesostructure details at the level of skin pores and fine creases to a resolution of up to a tenth of a millimeter. By recording a sequence of such scans or performing blendshape animation using scans of different high-res expressions, the effects of dynamic mesostructure - pore stretching and skin furrowing - can be recorded and reproduced on a digital character.

Recently, recorded skin microstructure at a level of detail below a tenth of a millimeter for sets of skin patches on a face, and showed that texture synthesis could be used to increase the resolution of a mesostructure-resolution facial scan to one with microstructure everywhere. They demonstrated that skin microstructure makes a significant difference in the appearance of skin, as it gives rise to a face's characteristic pattern of spatially-varying surface roughness. However, they recorded skin microstructure only for static patches from neutral facial expressions, and did not record the dynamics of skin microstructure as skin stretches and compresses.

Figure 2: Three real forehead expressions (surprised, neutral, and perplexed) made by the same subject showing anisotropic deformations in microstructure.

Skin microstructure, however, is remarkably dynamic as a face makes different expressions. Fig. 2 shows a person's forehead as they make surprised, neutral, and angry expressions. In the neutral expression (center), the rough surface microstructure is relatively isotropic. When the brow is raised (left), there are not only mesostructure furrows but the microstructure also develops a pattern of horizontal ridges less than 0.1 mm across. In the perplexed expression (right), the knitted brow forms vertical anisotropic structures in its microstructure. Seen face to face or filmed in closeup, such dynamic microstructure is a noticeable aspect of human expression, and the anisotropic changes in surface roughness affect the appearance of specular highlights even from a distance.

Dynamic skin microstructure results from the epidermal skin layers being stretched and compressed by motion of the tissues underneath. Since the skin surface is relatively stiff, it develops a rough microstructure to effectively store a reserve of surface area to prevent rupturing when extended. Thus, parts of the skin which stretch and compress significantly (such as the forehead and around the eyes) are typically rougher than parts which are mostly static, such as the tip of the nose or the top of the head. When skin stretches, the microstructure flattens out and the surface appears less rough as the reserves of tissue are called into action. Under compression, the microstructure bunches up, creating micro-furrows which exhibit anisotropic roughness. Often, stretching in one dimension is accompanied by compression in the perpendicular direction to maintain the area of the surface or the volume of tissues below. A balloon provides a clear example of roughness changes under deformation: the surface is diffuse at first, and becomes shiny when inflated.

While it would be desirable to simulate these changes in appearance during facial animation, curent techniques do not record or simulate dynamic surface microstructure for facial animation. One reason scale: taking the facial surface to be 25cm X 25cm, recording facial shape at 10 micron resolution would require real-time Gigapixel imaging beyond the capabilities of today's camera arrays. And simulating a billion triangles of skin surface, let alone several billion tetrahedra of volume underneath, would be computationally very expensive using finite element techniques.

In this work, we approximate the first-order effects of dynamic skin microstructure by performing fast image processing on a highresolution skin microstructure displacement map obtained as in. Then, as the skin surface deforms, we blur the displacement map along the direction of stretching, and sharpen the displacement map along the direction of compression. On a modern GPU, this can be performed at interactive rates, even for facial skin microstructure at ten micron resolution. We determine the degree of blurring and sharpening by measuring in vivo surface microstructure of several skin patches under a range of stretching and compression, tabulating the changes in their surface normal distributions. We then choose the amount of blurring or sharpening to affect a similar change in surface normal distribution on the microstructure displacement map. While our technique falls short of simulating all the observable effects of dynamic microstructure, it produces measurement-based changes in surface roughness and anisotropic changes in surface microstructure orientation consistent with real skin deformation. For validation, we compare renderings using our technique to real photographs of faces making similar expressions.


We record the surface microstructure of various skin patches at 10 micron resolution with a setup similar to which uses a set of differently lit photos taken with polarized gradient illumination. The sample patches are scanned in different deformed states using the lighting apparatus with a custom stretching measuring device consisting of a caliper and a 3D printed stretching aperture. The aperture of the patch holder is set 8 mm for the neutral deformation state and is set 30 cm away from a Ximea machine vision camera which records monochrome 2048 by 2048 pixel resolution images with Nikon 105 mm macro lens at f/16, so that each pixel covers a 6 micron square of skin. The 16 polarized spherical lighting conditions allow the isolation and measurement of specular surface normals, resulting in a per-pixel surface normal map. We integrate the surface normal map to compute a displacement map and use a high pass filter to remove surface detail greater than the scale of a millimeter to remove surface bulging.

Figure 3: Texture-aligned surface normal (top) and displacement (bottom) maps of a skin patch under vertical compression and stretching. (a) full compression, (b) medium compression, (c) neutral, (d) medium stretching, and (e) full stretching.

Each skin patch, such as part of the forehead, cheek, or chin, is coupled to the caliper aperture using 3M double-sided adhesive tape, and each scan lasts about half a second. After performing the neutral scan, the calipers are narrowed by 0.8mm and the first compressed scan is taken; this continues until the skin inside the aperture buckles significantly. Then, the calipers are returned to neutral, and scans are taken with progressively increased stretching until the skin detaches from the double-stick tape. Fig. 3 shows a skin sample in five different states of strain. The calipers can be rotated to different angles, allowing the same patch of skin to be recorded in up to four different orientations, such as the forehead sample seen in Fig. 4.

Figure 4: Each column shows measured 8mm wide facial skin patches under different amounts of stretching and compression, with a histogram of the corresponding surface normal distributions shown to the right if each sample. in microstructure.

With skin patch data acquired, we now wish to characterize how surface microfacet distributions change under compression and stretching. After applying a denoising filter to the displacement maps to reduce camera noise, we create a histogram of the surface orientations observed across the skin patch under its range of strain. Several such histograms are visualized in Fig. 4 next to their corresponding skin samples, and can also be thought of the specular lobe which would reflect off the patch. As can be seen, stretched skin becomes anisotropically shinier in the direction of the stretch, and anisotropically rougher in the direction of compression. For some samples, such as the chin in Fig. 4(g,h), we observed some dependence on the stretching direction to the amount of change in normal distributions. However, we do not yet account for the effect of the stretching direction in our model.

The variance in x and y of the surface normal distribution quantify the degree of surface smoothing or roughening according to the amount of strain put on the sample. Again, stretched skin becomes shinier, and compressed skin becomes rougher.


Figure 5: A sampled skin patch is deformed with FEM which drives microstructure convolution, rendered with path tracing.
[HD Rendering]

Fig. 5 shows frames from a sequence of a 1cm wide digitized skin patch being deformed by an invisible probe. It uses a relatively low-resolution finite element volumetric mesh with 25,000 tetrahedra to simulate the mesostructure which in turns drives dynamic microstructure convolution. The neutral microstructure was recorded using the system in Fig. 1 (left) at 10 micron resolution from the forehead of a young adult male, and its microstructure is convolved with parameters fit to match its own surface normal distribution changes under deformation as described in Sec. 5. The rendering was made using the V-Ray package to simualte subsurface scattering. As seen in the accompanying video, the skin microstructure bunches up and flattens out as the surface deforms at a resolution much greater than the FEM simulation.

Fig. 1 highlights the effect of using no microgeometry, static microgeometry, and dynamic microgeometry simulated using displacement map convolution with a real-time rendering. Rendering only with 4K resolution mesostructure from a standard facial scan produces too polished an appearance at this scale. Adding static microstructure computed at 16K resolution using a texture synthesis technique increases visual detail but produces conflicting surface strain cues in the compressed and stretched areas. Convolving the static microstructure according to the surface strain using normal distribution curves from a related skin patch produces anisotropic skin microstructure consistent with the expression deformation and a more convincing sense of skin under tension.

Figure 6: Real-time rendering of the forehead reigion under compression with dynamic microstructure (right) compared to static microstructure (left).

Questions and Answers

What is microgeometry?
Human skin features can be divided into roughly three scales: macro, meso, and micro. On a face, "macroscale" features define the overall shape of the facial features. This includes things like the jaw line, nose, and eyebrows. "Mesoscale" features are on the order of a millimeter (~0.1 mm) and include features such as pores and fine wrinkles. "Microscale" features are an order of magnitude smaller than pores and fine wrinkles (~10 microns) and describe the very fine deviations inside pores and along wrinkles. The following report extensively describes the anatomy of human skin: Link

Why is microgeometry important? Is this "microfacet"?
In the theory of rough surface reflections, a surface is composed of microscopic faces called "microfacets", which are assumed to behave like perfect minuscule mirrors. Our scan, captured at sub 10 micron resolution, is still an order of magnitude larger than a "real" microfacet. However it provides noticeable effects on surface reflectance. When it is seen from far away such as in a portrait, it breaks up specular highlights, appearing as high frequency glints on the surface. Figure 1 (a) shows that rendering with only a 4K resolution mesostructure from a standard facial scan. This produces specular highlights that appear too dull in an extreme closeup. Adding the static microstructure computed at 16K resolution with the texture synthesis technique from [Graham et al. 2013] increases visual detail as in Figure 1 (b). For more details about static microgeometry, we encourage you to read our previous paper.

Is there a low budget way to approximate the static microgeometry?
While [Graham et al. 2013] used measurement-based texture synthesis to synthesize 16K static microgeometry, [von der Pahlen et al. 2014] demonstrated that the effects of static microstructure can be approximated by a procedural noise in a real-time character demo Link.

What is the difference between this paper and the [Graham et al. 2013] paper?
While [Graham et al. 2013] focuses on the effect of static microstructures, our work investigates the dynamic appearance of the skin microstructures. Though adding the static microstructure improves the skin like quality, it produces conflicting surface strain cues when the skin is deformed significantly (Figure 1 (b)). On the other hand, our technique produces anisotropic skin structures consistent with the expression, providing a more convincing sense of skin under tension (Figure 1 (c)).

What happens when the skin deforms?
Generally speaking, the skin becomes rougher when it is compressed and smoother when it is stretched. Figure 4 shows that under deformation the skin surface normal distribution histogram (which can be viewed as a resulting BRDF) exhibits the roughness and anisotropy changes in a predictable manner.

How do you simulate dynamic microgeometry?
We approximate the skin being flattened under stretching, and bunched up under compressions by convolving a 16K displacement map. We blur the microgeometry displacement map in the direction of stretching, and sharpen it in the direction of compression using the surface normal distribution histogram as a guide. This entire computation can be efficiently implemented on GPU shaders.

Can I use dynamic microgeometry for realtime/offline applications?
The paper and the companion video include both realtime renders done with GPU shaders such as Figure 1, and offline renders such as Figure 5 skin slab.

How is the computed microstructure used together with the mesoscale structure?
In this paper, we used unconstrained blending, meaning that we simply added the computed 16K dynamic microgeometry displacement to the existing 4K mesogeometry displacement. If desired, constrained blending may be done by first converting displacement into a tangent normal map, and leveraging a normal map blending technique as done in [von der Pahlen et al. 2014].

Can I use physics simulation with this technique?
The microgeometry simulation framework allows deformation from any source, including physics simulation, keyframe animation, or captured facial performance as shown in this paper.


The authors would like to thank Randal Hill, Kimberly Lu, Ari Shapiro, Cary Peng, Bill Phelps, Emily O'Brien, Jay Busch, Xueming Yu, Etienne Danvoye, Javier von del Pahlen, the Digital Human League, Valerie Dauphin, and Kathleen Haase for their assistance and support. This research was sponsored by the U.S. Army Research Laboratory (ARL), the Funai Foundation for Information Technology, and in part by the National Science Foundation (CAREER-1055035, IIS-1422869) and the Sloan Foundation, and Royal Society Wolfson Research Merit Award. The content of the information does not necessarily reflect the position or the policy of the US Government, and no official endorsement should be inferred.


SIGGRAPH 2014 Talk
SkinStretch_SIG14t.pdf, (8.47MB)

SIGGRAPH 2015 Paper & Presentation
SkinStretch_SIG15p.pdf, (76.1MB)
SkinStretch_SIGGRAPH_Web.pptx, (358MB)

SIGGRAPH 2014 Video
SkinStretchTalk_SIGGRAPH2014.wmv, (167MB)

SIGGRAPH 2015 Video
SkinStretchPaper_SIGGRAPH2015.mp4, (191MB)
SkinStretchPaper_SIGGRAPH2015.wmv, (302MB)
EmilyReunion_2015CAF_1920x1080_30fps_v16.mp4, (65.2MB)

Sample Patch Data, (502MB)

GLSL Shaders
License and download

Related Projects

Measurement-Based Synthesis of Facial Microgeometry, SIGGRAPH 2012, Eurographics 201