Realtime Recognition of Handwritten Lontara Makassar Characters Using Yolov5 Algorithm
Isra Andika Bakhri, Hari Purnomo Sidik · 2022
Nowadays, handwriting pattern recognition has developed rapidly, including handwritten letters for ancient characters. The machine can be trained to read each letter. The use of lontara letters in South Sulawesi still exists today. To introduce lontara letters interactively, the author tries to research real-time reading of handwritten lontara letters to support these expectations. The results of this study succeeded in making this and getting satisfactory results with the evaluation value of precision rate (P) of 95.2%, recall rate (R) of 95.2%, and mean average precision (mAP) of 95.9%.