Award: www.2023.icsest.net/pages/29/...
For the organization and understanding of the continent's diverse heritage, the classification of African artifacts is of the utmost importance. The goal of this study is to use Machine Learning to create a classification system for the three main categories of African artifacts: clay, metal, and wood. The suggested system will help these artifacts be categorized, examined, authenticated, and conserved effectively. A total of 880 collected datasets were obtained from Google images with keywords specific to distinguishing between clay, metal, and wood. The four CNN models were used, and each model's performance was assessed to evaluate how effectively it classified the three different categories of artifacts. LeNet distinguished between items made of clay, wood, and metal with the best degree of accuracy (74%). Inception and Pretrained_Inception demonstrated their ability to classify artifacts with intermediate accuracy scores of about 61% and 65%, respectively. ResNet50, however, only managed to differentiate the artifact types with an accuracy of 58%, demonstrating its limits.
Proceedings: www.istes.org/proceedings-of-...
Paper: www.researchgate.net/profile/...