To Stylize or not to Stylize? The Effect of Shapes and Material Stylization on the Perception of Computer-generated Faces
To Stylize or not to Stylize? The Effect of Shapes and Material Stylization on the Perception of Computer-generated Faces
Eduard Zell, Carlos Aliaga, Adrian Jarabo, Katja Zibrek, Diego Guierrez, Rachel McDonnell, Mario Botsch
Bielefeld University, UniversidaddeZaragoza, Trinity College Dublin
ACM
Abstract
page link
Yatani Format
Motivation
What issues or problems did they try to solve?
Designing believable characters remains a challenging task. They wanted to analyze two of the most influential factors for a character's looks.
What was lacking in existing research?
Some works have already investigated the role of few of the many variables involved in the creation of 3D faces, such as rendering style, anthropomorphism or applying 2D filters. But shapes and materials have not yet been researched which affects 3D characters.
Method
What system did they create? Why did they think that the system design was well justified?
1a. They first investigated the influence of shape and material. They used three types of shapes (cartoon, middle, realistic) and three types of materials (cartoon, middle, realistic).
1b. They investigated shading and lighting by modifying the lighting transparency.
1c. They investigated the effect of pigmentation variation filtering through the gaussian blur of the kernel. 2. They added a different agender character for the experiment. They set 5 stylizations times 5 material levels times 5 expressions(surprise, angery, sad, happy and neutral). 3. They investigated the effect of shape on the recognition and intensity of the expression.
What kind of surveys did they do? How did they justify their study design?
From McDonnell et al. 2012 , they employed the following scales. (Appeal, Eerie, Relstic, Familiar, Attractive)
Much of the information that people use to evaluate virtual characters is available in a still image[McDonnell et al.2012]
What experiments did they conduct? How did they justify their experimental design?
The user’s task and the rating scales are explained before the experiment All stimuli were presented in a random order and shown for 3 seconds each. Participants were asked to rate stimuli according to the above scales after each stimulus presentation.
Insight
What results did they obtain? Under what conditions did it work, and under what conditions did it not work?
1a. Appeal and Reassurance were obtained with good results in toon and middle. 1b. The difference between realism and toon was very small in lighting. 1c. Blurred texture affected positive in Realism and Appeal with Realistic texture.
- They obtained good results in level2 and level3.
- In effect of shape on recognition, only the angry expression did not work. And the effect of intensity, the more shape was being tooned, the less intensity they obtained.
What are the new findings? What are the findings that could be used in other applications or systems?
Shape is the key attribute for perceived realism, however realistic materials do not increase realism of stylized shapes. But appeal, erriness and attractiveness are mainly affected by material, especially albedo texture is the dominant factor.
Contribution Summary
Eduard analyzed the relationship between shape and material in 3D characters and found shape is the main factor for realism and material is key to appeal, erriness and attractiveness.
