Pengenalan Tulisan Tangan Aksara Incung Menggunakan ResNet-50 Berbasis Mobile Application

Abstract

Incung writing is one of the traditional writings from the people of kerinci and sungai penuh, jambi. This writing has been used from the 14th century, as time goes on, the usage and understanding of this writing has gone down, especially for the younger generation. Because of the issue, a system is needed to support the preservation of the writing through optimized digitalization. This research develops a Handwritten Incung Characters Recognition system using the resnet-50 algorithm which will be android based. The dataset that will be utilized consists of 11.760 images that are divided into 28 classes. Testing shows that the resnet-50 algorithm has the capability to recognize the incung writing with the high-accuracy of 99% on training data, and 98% on testing data. This system has also worked when implemented on a mobile based application that detects words on real-time using the smartphone camera. According to results mentioned, it can be concluded that the resnet-50 algorithm has the ability to recognize the incung writing with great results.