Hair salon
Hair salon system that animates and renders hair at interactive rates. Because of the performance requirement, most interactive hair modeling algorithms do not support important, complicated features of hair, including hair interactions, dynamic clustering of hair strands, and intricate self-shadowing effects.
Most of the time, real hair appearance and behavior are sacrificed for real-time interaction with hair. Since the shape of the hair model and other properties are mostly dictated by hairstyling, it is an important step in modeling hair. Virtual environments created for interactive hairstyling are becoming key methods for understanding and specifying these hair properties and are useful for several applications including cosmetic prototyping, the entertainment industry, and cosmetologist training.
Accurate virtual hairstyling requires both high-performance simulation and realistic rendering to allow interactive use and the incorporation of fine details. Hairstyles result from the physical properties of hair and hair’s mutual interactions. Thus, hair salon dynamics need to be integrated to imitate the technique of actual global coiffure creation. Moreover, in the real world people are accustomed to hairstyling by touching hair directly. We employ a commercially available haptic device to present an intuitive 3D user interface. In this way, users can manipulate the hair salon in direct interaction with the 3D model, as in real-world hairstyling.
Interactive Hairstyling
In most hairstyling systems, static hairs are manipulated. One widely used technique applies clusters or wisps which couple strands of hair into groups, typically in a cylindrical shape, to accelerate the styling process. Generally in these methods, a curve defines the shape of the cluster and the user manipulates the shape of the curve to attain a desired hairstyle. Kim and Neumann prolonged the usage of generalized cylinders right into a multi-decision manipulated shape permitting the consumer to edit the hair salon geometry at worldwide and neighborhood levels.
Another approach is to use fluid or vector fields to control the direction of hair strands. Hadap and Magnenat-Thalmann used streamlines of fluid flow to create static hair shapes. Because of the performance requirement, most interactive hair salon modeling algorithms do not support important, complicated features of hair, including hair interactions, hairstyle, dynamic clustering, hair strands, and intricate self-shadowing effects.
Static 3D vector fields were used by Yu to model curly hairstyles. Recently, Choe and Ko published a method that uses vector fields and trajectories to make complex hairstyles. The user-defined trajectories constrain the global positions of hair salon strands while extra parameters are provided for defining curliness or randomness within a group of strands to provide localized detail. Hair shapes, including ringlet curls, have also been modeled using the natural curliness of hair.
Realistic Hair Lighting
We model the hair strand reflections by anisotropic light scattering from a cylindrical surface as proposed by Kajiya and Kay, augmented by recent observations made by Marschner, Jensen, Cammarano, Worley, and Hanrahan that one must also account for the multiple modes of scattering that occur inside and on the surface of the cylinders.
They demonstrated that a considerable amount of colored reflection arises from diffuse scattering, whereas most come from uncolored, specular reflection. We included two distinct specular highlights since there are multiple modes of scattering inside and on the surface of the hair fibers. The primary highlight is highly specular and shifts a little more toward the hair root the secondary highlight has a wider falloff and shifts toward the hair salon tip.
Realistic effects of self-shadowing within hair salon strands are difficult to implement very efficiently because the amount of dynamic geometry is large, and the hair salon strands are extremely fine. Regular shadow maps fail because the high frequency in the geometry forces us to use sampling rates beyond realistic values. We make use of two recently introduced features of graphics hardware to achieve real-time self-shadows. The first, multiple render targets, allows us to output 16 floating point values in one rendering pass. The second, floating point blending, makes high-precision blending of shadow values possible.
Haircutting
It is essential to cut hair when changing a hairstyle. Traditionally, the approach towards specifying hair length has often been to control the hair length through a parameter that either makes the hair grow or shorten. Some of the works have allowed local changes in the length parameter through either curve drawing, or through a 2D length map that correlates with the distribution of strands on the scale.
Alternatively, Lee and Ko cut hair with a cutting surface hair salon that intersect the surface are clipped to the surface shape. Our cutting method builds on the words of Lee and Ko to model the cutting of hair performed with scissors as is used in a salon. We version all of the functions of the reduce, which includes shooting the hair salon that falls away or is reduced. The cutting location is determined by a triangle created by the open blades of scissors.
Hair skeletons that intersect this triangle are cut. At the cutting location, the skeleton S is split into two separate skeletons, S1 and S2, S1 remains attached to the scalp, while the latter, S2, falls away. Two new control points are created at the intersection of skeleton S and the cutting triangle. Before the cut is performed, the geometry of the fallen hair salon remains consistent with the geometry of the hair below the sever point curliness, wetness, hair distribution, hair design, stylish cutting, and other properties are maintained in the fallen hair segments.
Hair Dryer
The strong force is communicated to any control points that are inside the cone of influence. In addition, the amount of water the wet control point will lose depends on the strength of the hair dryer force and decreases as the length of exposure with the hair dryer is covered. The hair dryer also can break the dynamic bonds created by hair spray.
Conclusion
Our virtual hair salon system has enabled physically based user interaction with dynamic hair salon while modeling several common hair salon applications. Level-of-detail representations coupled with our simulation localization scheme have accelerated the animation of hair so that a user can interact with it.
We presented a system involving an intuitive 3D user interface and methods for animating and rendering hair that allows a user to interactively manipulate hair salon through several common hair salon applications. This system provides a level of user interaction that has before been too complex to achieve. We are interested in experimenting with two-handed haptic gloves with our system in order to achieve higher fidelity force feedback on the part of the user, allowing further interaction capability and the creation of even more complex hairstyles.