Importance Of Handwritten Character Recognition

1155 Words5 Pages
Recognition of handwritten character is one of the most interesting topics in pattern recognition. The development of this type system can make better man-machine interface. The use of conventional keyboard for the interaction with the computer is problematic due to the large alphabet size in the case of Indic scripts. The combination of keys has to remember to input a character. Hence, handwriting recognition acquires great significance in the context of Indic scripts. Handwritten Character Recognition (HCR) means to classify handwritten characters into appropriate classes based on the features extracted from each character. Handwritten character recognition system can be of two type either online or offline. The image of handwritten text…show more content…
This system is for online HCR for Gujarati language. Data collection is carried out with the development of writing tool which collects writer’s input data by collecting sequence of co-ordinates points of the moving mouse or finger on touch pad. The writers were belonging to different age group as well as different educational background. They consisted of university students, school students, professors and employees in private companies. First dataset with 3700 samples for 37 characters of Gujarati language has been generated for recognition of character; second dataset with 1000 samples for 10 digits has been generated for recognition of…show more content…
In this stage, each preprocessed sample is transformed into a sequence of feature vectors. This paper describes the stroke based features. 12 low level strokes features and 8 directional features are extracted during the feature extraction. The prepressed sample of character is a combination of strokes. The division of character in to strokes in done. In first approach stroke is defined as the traced by the pen from a pen-down event to a pen-up event. It is represented using the data captured as the stroke is written. The number of points collected varies with the stroke and the approach and figure (2) shows the division of character done by the second approach. There are three steps for classification of characters namely low level stroke features and directional features extraction, feature vector generation and classification. Stroke based feature extraction is used in this work which include the extraction of 12 low level strokes. Different low level strokes are identified that represents basic elements of geometry namely end points (4), lines (4), and curves (4). The first feature vector is generated from these 12 LLS features. Hierarchical histograms up to fifth level are generated for these 12

More about Importance Of Handwritten Character Recognition

Open Document