Speech And Image Recognition Technology For Embedded Application

The capability of pattern recognition is an important part of human intelligence. Realizing the automatic pattern recognition by computer is a key breakthrough of intelligence machine developing. Pattern recognition is not only a research field having great scientific value, but also a great key technique which we need to break through in many applications of the digital network times.The speech and image recognition technology is an important developing direction of pattern recognition. With the improvement of people's living standard ,to day the computer is growing in popularity in China .In order to getting a better communication between computer and man, the key technology who relied on is the speech and image recognition ,which will become the significant technology in the information industry and be the important aspect of the future computer .At present ,the trend is using the speech and image recognition technology in embedded computer area .For the few products among people's daily life, the field should have a wide market prospects.Against this background, and based on complete and in-depth study in the speech recognition technology which combined with the image recognition, this paper achieves the speech and image recognition control to a toy robot.The paper's main finding is listed below:(1)Firstly ,it introduces the existing situation and a brief explanation of the fundamental and main methods of the speech and image recognition area.It brings a good foundation for further study.(2)After researching the fundamental principle of statistical pattern recognition and a summary in the recent preprocessing speech recognition technology,it pays attention to the analyses and comparison of the preprocessing technology and integrates the advantage of them and gives necessary instruction to realize the system.(3)According to analyse the way of speech feature parameter extraction ,the paper gives the principle and also studies the model-training and model-matching theory.(4)It analyzed the Image Segmentation technology and focused in introducing The maximum variance threshold segmentation under the subject requirement of this research.(5)It designs a system which is suited to speaker-independent and small-scale vocabulary speech recognition based on the SPCE061A controller ,and combined with Sunplus Technology Co.'s eagle image recognition module.The conclusion of this paper analyses the problems in recent research and points out the direction of subsequent research work.