Sign Language Fusion Platform Based On Software Reuse Technology

Sing language research can help the healthy people and deaf people to communicate, also can promote the development of anthropomorphic human-computer interaction. Sign language is usually expressed through gestures, facial expressions, head pose and any other means, but only hand information of sign language is not able to completely express the sign language information, the recognition rate is low, therefore multiple channels of sign language information is needed to fuse to improve the accuracy of sign language recognition.Sign language information fusion research involves many steps, each step will use a variety of different algorithms, so these algorithms need a platform to integrate them. The research platform which manages the various algorithms can avoid the system building repeated, at the same time, it also can realize resources sharing. This paper designed and realized a multi-channel sign language information fusion platform. This platform can quick and convenient to add the related algorithm of the various steps, and it also has a flexible interface, strong scalability, and the components in it can be shared.In detail, the main research can be described as follows:1. This paper constructed a development process model of integration system base on the field of sign language recognition and sign language. The FODA(Feature Oriented Domain Analysis) is applied to execute the information analysis, characteristics analysis and operation analysis for the field of sign language recognition and fusion recognition from the domain analysis view. By determining the boundary of sign language recognition and multi-channel sign language fusion recognition, the analysis model is gotten. For the variability of elements and fixed elements of sign language recognition domain, characteristics of domain component and system framework is designed. The multi-channel sign language information fusion platform which is based on domain engineering has established the basic function and the essential basis for the platform, and it also provides support to the reusability of the platform.2. Component reuse is applied to multi-channel sign language integration platform, the algorithms are encapsulated as the component and integrated into the platform, the platform also provides the component library management module with realizing the effective management and search of components.3. The design pattern is applied to design the multi-channel sign language platform’s architecture. Following the principle of object-oriented, in view of the various problems encountered in system designing, there are many design pattern means being used, for example, the mediator pattern, facade pattern, factory method pattern and strategy pattern. On this basis, it improves the reusability of the code.