Classification of Rincon Romaine Lettuce Using convolutional neural networks (CNN)
Angelo John F. Tenorio, Jerome Martin H. Desiderio, Cyrel O. Manlises
Abstract
Lettuce is in demand almost everywhere because it is the main ingredient in salads and cuisine. It has many variants, one of which is Rincon romaine lettuce. This variant is ideal for the hot and tropical temperatures of the Philippines because of its characteristics. The researchers want to have better plant survivability in the Tropics. In this research, the main purpose is to determine whether the sample is a Rincon romaine lettuce variant or not. The system or model used in this research is called Convolutional Neural Network, an ideal model for analyzing and processing fixed data. The researchers used a Raspberry Pi with a camera for image scanning and an LCD module for the input and output controls of the system. One of the objectives was able to prove the reliability of the system with the use of a confusion matrix. Using the allocated samples, the researchers were able to achieve an accuracy of 86.67%. The researchers have concluded all the research objectives with minimal errors or misidentification of the samples. The system may be beneficial to the producers in differentiating Rincon romaine lettuce from the other types and distinguishing the said lettuce when sourcing from different marketplaces.