Handwritten character and pictogram recognition for AV remote controller

This paper proposes a handwritten character and pictogram recognition system using a touch panel for data entry. The system consists of three processes: character entry on the touch panel; LCD indication; and pictogram recognition. For pictogram recognition, a handwritten character recognition engine, the free software Zinnia, was used. Then, after integrating this system of handwritten pictogram recognition into the μT-Engine, its action was verified. An experiment was also performed using human subjects to measure the recognition accuracy of this system. In the experiment, each of the 11 subjects wrote 92 characters, 52 upper and lower case alphabetic characters, 10 numerals, 0 to 9, and 30 symbolic pictograms in order. The experiment resulted in the recognition rate of this system being approximately 73% and the recognition rate of the pictograms being approximately 78%.