Kannada Character Recognition Using Multi-Class SVM Method

Character recognition also truncated to OCR (Optical Character Recognition), is the translation of images: handwritten, typewritten (either mechanically or electronically), or even just a simple text printed into a machine-editable text. Among them, one of the most relevant types is handwritten character recognition. Each handwritten content is composed of symbols, alphabets, etc. that are very syllabic with a distinct font style. Character recognition is achieved through segmentation, feature extraction, and classification, by using any one the machine learning methods thereby making tremendous advancements. This paper keenly focuses on the recognition of handwritten Kannada alphabets. Different machine learning classification strategies have known to be applied to achieve this recognition. Although, in this paper, we mostly center around the procedures dependency on the solution provided by Support Vector Machine(SVM) classifiers using Python. For simplicity, we focus on only four Kannada alphabets in this paper.