Model Descriptions for Vision and Graphics (コンピュ-タビジョンとコンピュ-タグラフィックス論文<特集>)

スポンサーリンク

概要

• 論文の詳細を見る

• All shapes can be represented as deformations from a standard or prototypical shape; it is thought that this is how shape is represented in human perception. An object's modes are the eigenvectors of its stiffness matrix, and are a physically-motivated way to obtain a canonical description of shape in terms of deformation from a prototype. Modes provide an efficient and reliable method for recovering, recognizing, and tracking a 3-D solid models from 2-D and 3-D measurements. They also provide an efficient, low-complexity method for computing complex, muli-body physical simulations. Several examples using this technology to model, recognize and track people will be presented. Software implementing many of these algorithms is available by anonymous FTP from whitechapel. media.mit.edu in the file /usr/ftp/misc/modal.tar/Z

• 1993-08-25

著者

• Pentland Alex

  The Author Is With The Media Laboratory.

関連論文

• Model Descriptions for Vision and Graphics (コンピュ-タビジョンとコンピュ-タグラフィックス論文)

スポンサーリンク