Tamil Characters Classification Using Random Kitchen Sink Algorithm (RKS)
R. Karthieswari, M. Sreethivya, Durgesh Kumar, K. P. Soman · 2014
Tamil is a classical south Indian language in India that contains more letters among other Indian languages. The motivation behind our work was to develop an efficient character recognition technique for analyzing the scanned Tamil documents. Initially character recognition was done with the help of frequently used Tamil characters. Based on that, we have created a new datasets by the process of segmentation using levelsets, binarization, and skew correction. In this paper we have recognized the Tamil characters by generating the random features using Random Kitchen Sinks. It gives the accuracy about 98.7%.