“Model-Based Analysis and Synthesis of Human Body Images” by Hoshino
Conference:
Type(s):
Interest Area:
- Application
Title:
- Model-Based Analysis and Synthesis of Human Body Images
Session/Category Title: Human Capture
Presenter(s)/Author(s):
Abstract:
Generating digital actors is important for many applications such as special effects in movies, games, and virtual reality. In this sketch, we present a mode-based analysis and synthesis of human body images. First, we estimate the current 3D pose of the human figure by using the spatio-temporal analysis with kinematic constraints. Then, we store the intensity images of the human body as textures of the 3D body model. Such a human model can be used to generate multiple views, merge virtual objects, and change motion characteristics of human figures in video.
References:
1. Hoshino, J., Saito, H., & Yamamoto, M. (2000). Match moving technique for human body images. SIGGRAPH 2000 Conference Abstracts and Applications.