抄録
The performance of a personal computer has improved very much and also the price has become low. Accordingly, the opportunity that amateur or physically handicapped persons utilize computers is increasing. But the typing from a key board has so far been a main input method to a computer. This method is annoying for these people. We developed a voice input system to a computer. In this system we are using a voice recognition device on the market. The device is a word recognition system for a special speaker.
There are many cases when voice recognition fails among similar utterance words. We propose the improvement method of recognition score that utilizes a simple learning system. This learning system is evaluated by voice recognition experiments. From experimental results, we have been able to confirm that this learning system has an improved effect in voice recognition.
