Suson, Jennelyn C.
Non-cursive handwritten word recognizer using neural network / Jennelyn C. Suson. - viii, 135 leaves : 28 cm.
Computer print-out
Thesis (Master in Computer Science) -- Cebu Institute of Technology - University, March 2011.
Handwriting recognition is a challenging task for many real-world applications such as document authentication, form processing, postal address recognition, bank check recognition, and interpretation of historical documents. With this, researchers have put an enormous effort into developing various techniques for handwriting recognition. This study presents a non-cursive handwritten word recognition system using neural network. This system has the ability to receive handwritten word from an input device. This handwritten word is preprocessed, cropped and segmented, and then feed into the neural network of training, classification and recognition. Handwritten word capturing of the system is done using a mouse-an input device, and is sensed online, that is, automatic conversion of word as it is written. Kohonen Self-Organizing Map (Kohonen SOM) neural network and image processing are the two main techniques applied. This study ensures that the processed word image data is read exactly the same with the original non-cursive handwritten word. The results show a practical application of the Kohonen neural network and are to what uses this technology might have.
Optical character recognition devices.
Writing--Data processing.
621.39/9
Non-cursive handwritten word recognizer using neural network / Jennelyn C. Suson. - viii, 135 leaves : 28 cm.
Computer print-out
Thesis (Master in Computer Science) -- Cebu Institute of Technology - University, March 2011.
Handwriting recognition is a challenging task for many real-world applications such as document authentication, form processing, postal address recognition, bank check recognition, and interpretation of historical documents. With this, researchers have put an enormous effort into developing various techniques for handwriting recognition. This study presents a non-cursive handwritten word recognition system using neural network. This system has the ability to receive handwritten word from an input device. This handwritten word is preprocessed, cropped and segmented, and then feed into the neural network of training, classification and recognition. Handwritten word capturing of the system is done using a mouse-an input device, and is sensed online, that is, automatic conversion of word as it is written. Kohonen Self-Organizing Map (Kohonen SOM) neural network and image processing are the two main techniques applied. This study ensures that the processed word image data is read exactly the same with the original non-cursive handwritten word. The results show a practical application of the Kohonen neural network and are to what uses this technology might have.
Optical character recognition devices.
Writing--Data processing.
621.39/9