In our current system, the interface provides the recognition/translation results and allows a user to select sign regions manually or confirm sign regions extracted automatically. For example, a user can select a sign that he/she is interested to be translated directly if multiple signs have been detected; or in some unfavorable cases the automatic sign extraction algorithms may fail, but a user can still select a sign manually by circling the areas using pointing devices. This function also allows a user to obtain a translation for any part of a sign by selecting the parts where he/she is interested. The following figures are examples of interfaces: the left one is the interface for development, and the right one shows the  interface on an PDA.     

 

  Sign Recognition

  Sign Translation

  Sign Capture

  User Interface