Grant-in-Aid for Scientific Research (A)
|Allocation Type||Single-year Grants|
|Research Institution||Japan Advanced Institute of Science and Technology (JAIST)|
MIYAHARA Makoto JAIST,School of InformationScience, Professor, 情報科学研究科, 教授 (00115122)
HORITA Yuko Toyama University, Dept.Engr., Assistant Prof., 工学部, 講師 (80209303)
KOTANI Kazunori JAIST,School of Information Science, Associate Prof., 情報科学研究科, 助教授 (20225452)
HASHIMOTO Hideo Kanazawa university, Dept.Engr., Professor, 工学部, 教授 (00251934)
HIBINO Yasushi JAIST,School of Information Science, Professor, 情報科学研究科, 教授 (10251969)
HARASHIMA Hiroshi Tokyo University, Dept.Engr., Professor, 工学部, 教授 (60011201)
|Project Period (FY)
1995 – 1996
Completed(Fiscal Year 1996)
|Budget Amount *help
¥1,700,000 (Direct Cost : ¥1,700,000)
Fiscal Year 1996 : ¥1,700,000 (Direct Cost : ¥1,700,000)
|Keywords||Multi Media / High Quality Image / Deed Impressive Quality / Kansei Information / Reality / Picture Feature Evtraction / Picture Quality Exaluation|
1.Research of Extra High Quality Multi-Media Communication System and Evaluation for factors of High Quality
(1) Experimental Equipment of Extra High Quality Multi-Media Communication System
We have tried to product the 36 bits/pel Image Memory and Display system which has high speed data transmission function via networks. It has a high speed CPU so a high speed and flexible data processing which is a comparative evaluation and test patterns generation are available.
(2) Research of Factors for High Quality Imaging
We have shown the high quantizing accuracy makes a high level picture representation using 36 bits/pel Imaging System. We can perceive a depth or 3D feeling by the detail of a contrast. High quantizing accuracy is indispensable to Extra High Quality Media Representation because 36bits/pel system has a high quality image representation comparable with photo or printing quality.
(3) Research of Factors for Deeply Impressive Quality Sound
We analyzed evaluation words participate wit
h high quality sound representation using statistical analysis. As the result, the representativeness of a high level sound feeling, i.e.profound impression of sound and sense of holographic depth of sound etc., are very important. The physical factors, i.e.representativeness of wave pattern, phase, group delay, etc., are participate with the high level sound feeling.
(4) Research of Picture Quality Scale
We developed a objective picture quality scale (PQS) for evaluation of coded pictures in the case of 1) mono chrome moving picture and 2) color still picture. PQS is able to estimate the Mean Opinion Score with about 90% accuracy using distortion factors,
・factors for mono chrome moving picture : smoothness and naturalness of moving
・factors for color still pictures : degradation of a structure of image contour and
continuousness of image structure for Hue, Chroma and Value components in the uniform color space.
2.Research of Media Synchronization
We developed an analysis and recognization of human facial expression using facial parts (i.e.extraction of eyes, nose and mouse) by Fractal analysis. Using Fractal analysis, it can extract each parts without influenced to a condition of picture capturing. Also the system can recognize facial expression in the case of large and typical facial seens (i.e.laughing, yawn, ...). Less