Revolutionizing Pistachio Detection: An Advanced VGG-16 Neural Network Approach

This research paper study suggests the classification of pistachio images using a deep learning-based approach, one of the features where VGG-16 architecture is utilized. The dataset that has been collected using Kaggle owns 2148 images with a split of training (80%) and validation set (20%). In order to get the best performance out of the model, various preprocessing techniques, such as image resizing and normalization, were applied. Precision, recall, and F1-Score values were 0.9671, 0.9666, 0.9671, respectively with accuracy of 0.9674 on the VGG-16 architecture. Training was performed for 50 epochs using Adam optimizer and categorical cross entropy loss; validating accuracy to be above 95% by the 20th epoch. Key findings on the model are that it converges very quickly and has high performance with significantly better performance than other state of the art methods in pistachio classification. The analysis of the confusion matrix confirmed its precise capabilities in classification. In addition to that, the VGG-16 architecture proved to have good potential application in real-world industrial settings where semi-autonomous systems benefit from high accuracy within relatively efficient computation. This research clarifies viability for the use of deep learning models in agriculture, delineating a scalable and efficient solution for pistachio classification and potentially extending to other agricultural products.