Loading...
Please wait, while we are loading the content...
Similar Documents
AniME : Automatically Generating Anime Portraits
| Content Provider | Semantic Scholar |
|---|---|
| Author | Nguyen, Michelle Li, Jingyi |
| Copyright Year | 2016 |
| Abstract | The popularity of anime (Japanese animated productions) has transcended cultural boundaries, resulting in high volumes of fan works such as personalized avatars. However, drawing aesthetically pleasing anime is a craft that takes a long time to develop. AniME (pronounced ani-mee) is a pipeline which integrates computer vision, computational photography, and data-driven frameworks to automatically generate an animestyle portrait from a user given photo. Our approach renders the corresponding lineart using landmark facial annotations, while features such as eye shape and hairstyle are obtained via template matching to a hand-drawn library of anime eyes and hairstyles. Finally, colors of the anime portraiture are determined through simple averaging and clustering of selected image patches. We validate our tool using photos of people of varying genders, ethnicity, and hair styles. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://jingyi.me/static/anime.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
