PENGENALAN TULISAN ARAB MENGGUNAKAN UNITED MOMENT INVARIANT DAN MODIFIED QUADRATIC DISCRIMINANT FUNCTION
Daftar Isi:
- Arabic handwriting has a unique number of letters that are almost similar but not identical. The writing style of each person is different in writing Arabic letters so that it becomes a distinguishing factor in the recognition of Arabic writing. The recognition of Arabic handwriting in this study uses United Moment Invariant (UMI) to extract features and Modified Quadratic Discriminant Function (MQDF) as an recognition method. Training data were 560 data and 224 test data in the form of PNG format images with a resolution of 30x30 pixels, the study resulted in accuracy reaching 87.05% in learning rate 0.6 with a processing time of 14 minutes 28 seconds. At learning rate 0.4 the accuracy results reach 88.39% with a processing time of 15 minutes 16 seconds.