Paper Title
Cacao Pod Variety Classification using ResNet Modelled Convolutional Neural Network

Abstract
Not being able to self-pollinate, cacao trees use cross-pollination as its primary solution. However, being sensitive to cross-pollination, it has no means of identifying its variety while growing. It causes confusion on the recognition of the diverse varieties of the pods. Traditional methods require identification of the seed before being planted, or identification after the cacao pods were harvested. Nonetheless, it projects negative effects in the cacao industry as there is poor quality control, having all the types of the pods labelled under the same category in pricing due to insufficient information regarding its varieties. With this, the proponents aim to develop a non-invasive cacao pod variety classifier using ResNet Model Convolutional Neural Network. The model was trained to identify three cacao varieties: Criollo, W-10 and BR-25. An image of the pod is to be taken as input for the model. The results show that the model achieved a precision of 90% and an accuracy of 95% over a total of 60 trials. Keywords - Cacao Pod, Neural Network, Residual Network (ResNet), Variety Classification.