Digital Culinary Evolution by Leveraging CNN’s for Pandang Cuisine Enhancement

Through the use of Deep Convolutional Neural Networks (CNNs) and Transfer Learning, the Pandang Cuisine Classification topic is investigated in this study. AI is being utilised to improve the efficiency and profitability of the food and beverage industry. A total of 993 photographs of nine different Pandang meals were used in the research project, which acknowledged Industry 4.0. Before training the model using Model Checkpoint and Early Stopping callbacks, the data is imported, converted, and pre-processed using the CNN model that has already been pre-trained using VGG19. The hyperparameters consist of two hundred epochs, thirtytwo batches, and a nine-class output layer. When precision, recall, and F1 score are taken into consideration, the model achieves an accuracy of $90 %$ on the test dataset. Accuracy and loss curves, test data projections, Classification Reports, Confusion Matrix, and Grad-Cam visualisations are included in the conclusion of the study. The findings of this study offer suggestions for methods to enhance the management of food supply chains and culinary operations.