Model for Handwritten Recognition Based on Artificial Intelligence
Narumol Chumuang, Mahasak Ketcham · 2018
This paper proposed a general algorithm for more efficient handwritten recognition. Using handwritten recognition algorithms can reduce the time it takes to convert documents into letters for reducing the workload. The handwritten fonts used in this thesis are multi-script, which consists of Bangla font, Latin, MNIST handwritten alphabet series on prescription. This step has been designed and developed with genetic algorithms in conjunction with artificial intelligence techniques. The result of this algorithm was designed and developed to produce accurate results in the recognition of the Bangla set is 94.05 %, Latin 98.58 %, and MNIST 100 %.