Implementation of Artificial Neural Network (ANN) in the Image Recognition of Offline Cursive Handwriting

Fitrianingsih Fitrianingsih (Gunadarma University)
Diana Tri Susetianingtias (Gunadarma University)
Dody Pernadi (Gunadarma University)
Eka Patriya (Gunadarma University)
Rini Arianty (Gunadarma University)

ILKOM Jurnal Ilmiah
Vol 14, No 1 (2022)

Identifying a writing is an easy thing to do for human, but this does not apply to computers, in particular if it is handwriting. Handwriting recognition, especially cursive handwriting is a research in the area of image processing and pattern matching that is challenging to complete, following the different characteristics of each person's cursive handwriting style. In this study, the use of the ANN model will be implemented in performing offline handwriting image recognition. The cursive handwriting image that has been obtained is then preprocessed and segmented using bounding box rectangle and contour techniques. Evaluation of system performance using global performance metrics in this study resulted in a percentage of 93% where the bounding box and contour succeeded in determining the segmentation point correctly, so that the ANN model worked optimally.