| Project/Area Number |
63420031
|
|---|
| Research Category |
Grant-in-Aid for General Scientific Research (A)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
電子通信系統工学
|
|---|
| Research Institution | The University of Tokyo |
|---|
Principal Investigator |
HARASHIMA H The Univ. of Tokyo, Dept. EE, Professor., 工学部, 教授 (60011201)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
AIZAWA K The Univ. of Tokyo, Dept. EE, Lecture., 工学部, 構師 (20192453)
TAZAKI R The Univ. of Tokyo, Dept. EE, Assistant., 工学部, 助手 (30143390)
SAITO T Kanagawa Univ., Dept. EE, Associate Professor., 工学部, 助教授 (10150749)
荒川 薫 東京大学, 工学部, 助手 (30183734)
|
|---|
| Project Period (FY) |
1988 – 1990
|
| Project Status |
Completed (Fiscal Year 1990)
|
| Budget Amount *help |
¥27,100,000 (Direct Cost: ¥27,100,000)
Fiscal Year 1990: ¥4,000,000 (Direct Cost: ¥4,000,000)
Fiscal Year 1989: ¥10,300,000 (Direct Cost: ¥10,300,000)
Fiscal Year 1988: ¥12,800,000 (Direct Cost: ¥12,800,000)
|
|---|
| Keywords | Future Image Communication / Intelligent Coding / Feature Extraction Coding / Analysis Synthesis Coding / Image Compression / 知的画像符合化 / 特徴抽出符合化 / 分析合成符合化 |
|---|
| Research Abstract |
This research investigates the basic technologies for the realization of intelligent image communications and the development of such new image coding algorithms as "feature extraction coding" and "model-based analysis-synthesis coding". The following list summarizes the main features of this work.
Model-based analysis-synthesis image coding. We have studied the major components of model-based coding : 3-D structural modeling of known objects, analysis and synthesis of image sequences.
Model-based/waveform hybrid coding. We designed a hybrid algorithm to integrate the model-based coding and waveform coding. The combination improves the generality and robustness of the coding system.
3D structure information from image sequences. We have developed robust algorithms for recovering the 3D information automatically. The recovery process specifically integrates cues that can be directly integrated into the model-based coding.
Analysis and synthesis of facial expression. The analysis and synthesis of facial expression are by far the most critical part of the model-based coding scheme. We have developed FACS-based and "eigenface" approaches.
Facial images from speech and text Model-based coding promises a broad range of applications. Among them is the conversion of media for usage in multimedia. Particularly We have concentrated in technigues that convert speech and text to dynamic facial images.
|
