Ricky Eka Putra, I Made Suartana, Rahadian Bisma, Miftakhul Jannah, Eka Cahya Maulidiyah, Anita Qoiriah
Education, which is the key to success, is often the most valuable inheritance from parents to children. Not infrequently parents are willing to spend a lot of money for the best education for their children. Early Childhood Education which is early education for children does not escape special attention for parents. Rapid child development begins at this stage. In the final era of the Covid-19 pandemic, digital learning applications are expected to be a bridge and a solution to improve children's education, especially in the preschool period. One of the abilities of early childhood that needs to be honed is the introduction of objects around them. This object socialization activity also needs to be supported by object recognition applications to help parents or teachers improve children's abilities. Therefore, we propose an Object Recognition System (OR-Sys) that is connected to a camera to help children recognize objects around them. The image recognition process of surrounding objects requires a reliable classification method. Convolutional Neural Network (CNN), which is the latest development of the Neural Network method, can adopt and improve the advantages possessed by previous methods, especially in terms of classification. In this study, we choose to use two CNN architectures that are quite popular in use, namely GoogLeNet and ResNet. Both architectures play an important role in the success of OR-Sys in carrying out its duties. This can be seen from the high level of accuracy generated by two architectures, which are 91,4% for GoogLeNet and 94,8% for ResNet. © 2023 IEEE.
Universitas Negeri Surabaya, Informatics Department, Surabaya, Indonesia; Universitas Negeri Surabaya, Information System Department, Surabaya, Indonesia; Universitas Negeri Surabaya, Department of Psychology, Surabaya, Indonesia; Universitas Negeri Surabaya, Early Childhood Education Department, Surabaya, Indonesia
