Algoritma Deep Learning Untuk Pengenalan Gambar Jenis Daun

Deep Learning Algorithm for Leaf Type Image Recognition

Abstract

Image processing is a branch of informatics that deals with transforming one image into another using certain techniques. Deep learning algorithms have become one of the effective approaches to solving this problem. In this paper, we propose a deep learning algorithm that uses Convolutional Neural Networks (CNN) architecture to recognize leaf types based on a given leaf image. We outline the main steps in model development, including data pre-processing, CNN architecture selection, and model training. The experimental results show that the proposed deep learning algorithm can achieve a high level of accuracy in leaf-type image recognition. In this study, the CNN method is used to identify and classify objects in digital images, specifically leaves. The dataset used consists of 33 leaf classes, with a division of 16,500 data for training, 3,300 for validation, and 1,650 for testing. The training and validation processes were carried out in as many as 150 epochs, which resulted in the highest accuracy of 94% with the lowest loss of 0.28. While in the testing process, the accuracy value obtained reached 84%. The researched method, which integrates CNN with data augmentation and transfer learning, demonstrated superior performance with an accuracy of 94% in leaf type recognition. This outperforms other methods that rely solely on traditional CNN or do not utilize augmentation and transfer learning, which generally achieve lower accuracy rates. The combination of these techniques enables more robust feature extraction and better generalization, leading to more accurate and reliable classification results compared to other approaches.