Accuracy Analysis on Images Retrieval System using Radial Basis Function Algorithm and Coefficient Correlation

• Published in: 2020 International Conference on Data Science, Artificial Intelligence, and Business Analytics (DATABIA) pp. 99 - 104
• Main Authors: Sinuraya, Khairul Abdi, Suwilo, Saib, Lydia, Maya Silvi
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.07.2020
• Subjects: Clustering algorithms; Coefficient Correlation; Correlation; Feature extraction; Image retrieval; Mathematical model; Radial Basis Function (RBF); Radial basis function networks
• DOI: 10.1109/DATABIA50434.2020.9190227

The image retrieval system is a system used for the process of retrieval of images based on information contained in the image files. Radial Basis Function (RBF) is one of the Neural Network methods used in the image retrieval system, is known for the capability to produce image information search properly. In determining the initial centroid value, the RBF method uses K-Means Clustering. This algorithm has a weakness in determining the right initial centroid value to get proper classification results in image retrieval. In this paper, the Coefficient Correlation (CC) method is used in determining the initial centroid value of the input data following the similarity of the data. Data with the highest degree of similarity compared to other data used as the initial centroid value. Data used in this study are leaf image data of 500 images with 10 categories of leaf types, and each sample contained 50 images. Based on the testing results, an increase in image retrieval accuracy with an average of 90.92% using the RBF and CC methods compared the image retrieval results using the RBF and K-Means Clustering methods gained an average accuracy of 85.96%.