VHDL described finger tracking system for real-time human-machine interaction

This work presents an image recognition technique that enables finger position to be automatically detected in natural interaction systems, within the context of ambient intelligence, AmI. The aim is a robust system in which users do not need to wear gloves containing specific colored markings to establish location. The system has been validated, having been used for the design of a virtual keyboard. The algorithm is based on the Hough transform to detect the presence of straight lines in a scene, as the fingerpsilas profile is sufficient for this method to be employed. To distinguish between a finger and other straight forms that may accidentally appear in a scene, the algorithm fine-tunes location by incorporating movement information. The system is implemented using a field programmable gate array (FPGA), providing real-time capability and improving the sensation of interaction for the user. In this article, we present a detailed description of all the stages involved.