The Design And Implementation Of Human-Computer Interaction Assisted System For Disabled Person

Speech recognition and computer vision technology promote the intelligent human-computer interaction technology continue to develop. Intelligent human-computer interaction technology not only bring ordinary users more convenience and efficiency, but also bring people with disabilities to change the nature, so that they can use computers more efficiently to improve their living conditions.Because the relatively low purchasing power of persons with disabilities, while have different needs in products, leading to high cost and less product of the corresponding human-computer interaction support system development. At the same time, as the current speech recognition and computer vision technology constraints, resulting in speech recognition and computer vision technology can't substitute for common using of mouse and keyboard. However, the speech recognition and computer vision can not meet the general needs of the user's intelligent human-computer interaction, as the disabled users, they desperately need these technologies to serve them. They can use the computer exchange with the outside world, to learn and improve, to enjoy the entertainment. This paper makes use of existing speech recognition and computer vision technology to develop a human-computer interaction support system for disabled users.Firstly, this paper designs and implements a Chinese language speech input system for persons with disabilities, improving the disabled persons especially with hand disabilities operating computer efficiency. The Chinese speech input system allows the users, can not use the keyboard and mouse, complete simulation of the standard keyboard and mouse action using only speech. Although the efficiency of system input is lower than mouse and keyboard input efficiency, as an alternative to low-cost way, this system can achieve a pure voice input of persons with disabilities support system. This system uses longer phrases as voice commands, so the system achieves relatively high recognition rate in low noise environment. After voice training, this system's recognition rate can reach 80%, and can maintain up to 70%recognition rate even in the noise environment.Subsequently, this paper implements energy-saving software based on combining face detection and differential detection. The software achieves shutting off the computer monitor power within 50 seconds when users is left. And this paper implements finer positioning based on image difference and corner detection. Mapping the coordinates of image and screen, this paper achieves the mouse movement and positioning analog functions. This paper detects eye pupil center in the infrared images using the edge detection and ellipse fitting, to support the research of eye-tracking system in the future.Finally, this paper finishes preemption threshold-priority real-time scheduling algorithm, after analyzing real-time methods in the embedded mobile devices which provide intelligent human-computer interaction, to support the intelligent computing mobile devices which have both real-time and low power consumption demands.