OHMI Tadahiro Tohoku University, New Industry Creation Hatchery Center, Professor, 未来科学技術共同研究センター, 教授 (20016463)
KOTANI Koji Tohoku University, Graduate School of Engineering, Associate Professor, 大学院・工学研究科, 助教授 (20250699)
|Budget Amount *help
¥51,090,000 (Direct Cost: ¥39,300,000、Indirect Cost: ¥11,790,000)
Fiscal Year 2002: ¥18,590,000 (Direct Cost: ¥14,300,000、Indirect Cost: ¥4,290,000)
Fiscal Year 2001: ¥32,500,000 (Direct Cost: ¥25,000,000、Indirect Cost: ¥7,500,000)
This research aims at realizing of personal mobile system which can instantly import and modify information in worldwide through global network, then make a transmission or a record, and quickly find and utilize a required datum in enormous amount of information. In order to realize such system, instant transmission of the high-definition picture information, record, and search are essential technologies. Therefore, we developed the algorithm which raises the rate of picture compression after separating object completely at the time of imaging, and built the foundation of realization of the intelligent man-machine-interface system equipped with recognition and judgment processing.
First, in order to form amorphous photoelectric film used as photoelectric conversion layer without structural failure, we created the gas flow control system featuring the gas pressure control method, which can instantly control composition of gas-mixture in film formation chamber, and microwave excited high-
density plasma equipment having dual shower-plate featuring low electron temperature, in which the novel exhaust pumps featuring non-equivalent pitch and non-equivalent angle of lead screw are implemented. By using this equipment, we established the foundation to form the highly reliable photoelectric film, which is formed by high-quality CVD, and highly selective RIE for miniaturized contact/ via-hole. Both of processes are realized by the selection of plasma-excitation gases and process-gases, respectively, which can realize suppression of both excess-decomposition of process gases and ion-bombardment-damages.
Moreover, in order to realize the highly sensitive and highly efficient image sensor which carries out noise control while imaging picture at the video rate and extracts the attributes and the amount of features which each imaged object has characteristic, such as a position of the depth direction, motion, color balance, and roughness (spatial frequency), on real time, and carries out category separation, we examined the device, circuit, and processing algorithm in detail, and completed the chip design.
Furthermore, we devised the technique of choosing the optimal image compression algorithm for every divided object category and established the foundation of the new vector quantization technique which maintains high image quality and realizes the overwhelming high compression rate. Less