<![CDATA[In the culinary and health fields, spices are known to play a significant role. However, differentiating and identifying spices can often be challenging, especially for students or individuals with impaired sense of smell and taste. Our research builds upon our previous study, which involved designing a spice classification model using the pre-trained EfficientNetB0 model, yielding a testing accuracy of 48%. This study designs and develops an application for classifying images of spices from the Zingiberaceae family using the Rapid Application Development (RAD) method, aimed at addressing the difficulties of manual spice identification. Testing results using a black box technique indicate that the designed application performs well in basic tasks such as opening the interface without significant errors, responding promptly to functional buttons, and displaying user-inputted image results. However, in one of the testing scenarios, the application struggled to accurately classify certain spice images, indicating the need for improvements in the model used for more precise spice image classification.]]>
Copyrights © 2024
