<![CDATA[One of human primary food source in the world is rice. As one of a primary food development regarding its types has been carried out, but due to the many types of rice that exist, it becomes difficult for consumers to tell a difference between one type of rice and another. Therefore, we need a solution to this problem. By using ResNet architecture on CNN algorithm and various data samples and test, it is hoped that this research can provide solutions to the problems encountered. The dataset consists of 75,000 image data divided into 5 types of rice which contains aborio, karacadağ, ipsala, jasmine, and basmati where the dataset will be divided in 80% data train to 20% data test ratio . The test scenario used in this study is a combination of types of optimization (ADAM, RMSProp, and SGD) and also types of pooling, namely max pooling and average pooling. After going through the test scenarios, the model with ADAM optimizer and Average Pooling gets the best accuracy of 98.62%.]]>
Copyrights © 2026
