Pengembangan Aplikasi Android Untuk Konversi Gambar ke Teks dengan Flutter dan OCR Menggunakan Metode Jaro-Winkler
Abstract
The need for tools for effective image to text conversion is increasing as a result of increasing reliance on mobile applications for daily activities and constraints in cross-platform application development. Addressing this demand by creating an Android application that uses Flutter and Optical Character Recognition (OCR) technology is the focus of this research. In particular, Jaro Winkler's method, which improves the accuracy of text extraction, is the main focus of this research. The research methodology uses a systematic approach, starting with a literature review to create a theoretical foundation and discover differences in current technology. The development uses Flutter's cross-platform capabilities and includes OCR for text extraction. The Jaro Winkler method is used to improve accuracy by increasing the similarity measure between extracted text and actual text, especially for unclear images. Despite issues with low-resolution images, the app has the ability to convert images to text accurately, as demonstrated by app testing and evaluation. The app's performance shows the possibility that a combination of Flutter, OCR, and Jaro Winkler's method can help create an effective image to text conversion tool. But the research also found some things that need improvement, especially in processing low-quality images and refining algorithms for better accuracy. Implementation of the Jaro-Winkler algorithm is used accurately and significantly. This is proven by a comparison of two sample words, "BRI" and "BNI," which produces a result of 0.7778, or a percentage value of 77.3% through manual calculation. Based on a comparison of the values in the application and calculations using Python with floating-point representation, there is a small difference in the final results, with a difference of only 3%, which results in a result of 80%..