Deteksi Spasi Antarkata Pada Tulisan Tangan Menggunakan Convolutional Neural Network Dengan Framework You Only Look Once Versi 11

Abstract

Advances in image processing and deep learning technology enable more accurate handwriting analysis, including the detection of interword spacing, which exhibits high complexity due to variations in writing styles. This study aims to implement a Convolutional Neural Network (CNN) algorithm using the You Only Look Once version 11 (YOLOv11) framework to detect and classify interword spacing zones into three classes: Narrow Word Spacing (NWS), Medium Word Spacing (MWS), and Wide Word Spacing (WWS). The dataset comprises 150 handwritten images with a total of 4.117 annotated interword spacing objects. The research methodology involves testing the model across variations of learning rates (0.1, 0.01, 0.001, and 0.0001) and data split ratios (70:30, 80:20, and 90:10). Model performance was evaluated using Precision, Recall, F1-Score, and mean Average Precision (mAP) metrics. Based on 12 experimental trials, the best configuration was achieved with a learning rate of 0.001 and a 90:10 data split. This configuration produced an mAP@50 of 0.455, an mAP@50–95 of 0.261, and an F1-Score of 0.49. These results indicate that the YOLOv11 model is capable of detecting interword spacing zones with reasonably good performance, despite remaining classification errors due to visual similarities between classes.