Co-Investigator(Kenkyū-buntansha) |
NIYADA Katsuyuki MATSUSITAGIKEN LTD., RESEARCH CENTER,PRESIDENT, ヒューマンインターフェイス研究所, 所長
MAKINO Shozo TOHOKU UNIVERSITY,COMPUTER CENTER,PROFESSOR, 大型計算機センター, 教授 (00089806)
SAGISAKA Kaoru TOHOKU UNIVERSITY,MEDICINE,PROFESSOR, 医学部, 教授 (70006740)
|
Budget Amount *help |
¥8,800,000 (Direct Cost: ¥8,800,000)
Fiscal Year 1997: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 1996: ¥1,500,000 (Direct Cost: ¥1,500,000)
Fiscal Year 1995: ¥6,500,000 (Direct Cost: ¥6,500,000)
|
Research Abstract |
The number of medico-legal autops cases is increaing with year in every part of Japan, while the preciseness of autopsy findings as evidence is demanded more strictly than ever. Therefore, the medico-legal practitioners cannot help spending much time on making the documents. Considering these circumstances, we developed a continuous speech input system in which oral statements of autopsy findings were converted into sentences. The characteristic of this study is that the forensic pathologist as the users connected closely with the computer scientists. In current speech input system, it is still difficult to automatically recognize continuously spoken general sentences. But, when the condition is so limited that a user is special, 1) a user is limited to a specified speaker, 2) the expression of document is almost decided, and the structure of sentence is comparatively simple, 3) the vocabulary size in the autopsy reports is 4000 from 3000, and the number of words in the vocabulary decreases more by dividing the document into each of physical parts. These conditions will lighten load to a device, and the device will be developed with ease. Firstly, we built an automaton by ECGI method to represent the structure of sentence. Then we defined the distance between the states of an automaton to calculate similarity of the words, the appearance of which was expected. Based on this definition, we developed a method in which an automaton was revised and generalized. Hiraiwa, Sagisaka and Makino were in charge of these development. For phoneme recognition, we used the model speech method developed by Niyata. We, all the members, put the above-mentioned methods together and made the speech input system of autopsy findings. The system ran to recognize sounds without time delay. Since the precision of speech recognition is not enough, the improvement of the system will be continued in future.
|