Optimized fingertip blob recognition for image processing based touch-screens

Almost undoubtedly, the greatest invention of the 20th century has been the computer. The computer uses several auxiliary devices to establish a seamless interaction with the user. One challenge faced during the development of these devices, was the lack of direct interfacing between the machine and the user. Today we see Natural User Interfaces (NUI), which include touch activated and speech activated inputs. One problem faced with the touch NUI's is ambiguity. This is mainly due to variations in finger size, proximity of icons, and the fact that the imprint of the finger, called a fingertip blob, may have to be mapped to a single pixel. In this paper we attempt to address the above problems, by focusing on achieving a high level of accuracy at low costs, utilizing image processing techniques, all the while keeping the process simple and fast. A novel algorithm for the same has also been discussed. The algorithm was implemented using the Java programming language on the Java Software Development Kit (JDK) version 1.5.0 (update 22). An accuracy of 100% in fingertip blob detection was achieved and an improvement of 15% was achieved in terms of execution time in comparison with existing algorithms, for various experiments conducted utilizing the new algorithm.