Edge Detection Algorithm and Image Classification in Digital Image Processing

Abstract

Digital image processing is one of the fields that is constantly evolving in tandem with artificial intelegence and machine learning. The purpose of this study is to perform a systematic review of various methods used in digital citra classification and detection, as well as to compare the effectiveness of traditional algorithms like Sobel, Prewitt, and Canny with deep learning-based algorithms like Convolutional Neural Networks (CNN). The Systematic Literature Review (SLR) method is used to identify, evaluate, and analyze 78 related articles that provide preliminary data wit inclusive and exclucive criteria from 300 articles. The primary focus includes edge detection, citra-based object classification (fruit, leaf, face, and medical image), and fitness entrapping techniques such as GLCM, LBP, HSV, and fraktal. The study's findings indicate that deep learning, specifically CNN, consistently provides higher accuracy in classification than traditional methods, despite requiring a larger computer power. It is hoped that this research would serve as a guide for developing more effective and efficient digital image processing systems in a variety of application domains, including pharmaceutical, medical, digital safety, and creative industries.