Classification of Regional Food Using Pre-Trained Transfer Learning Models

Using the most demanded technology computer vision and machine learning, image classification is the process of drawing meaning or comparable patterns from an image. Today, a transfer learning framework is used to perform image categorization tasks. Transfer learning is the process of increasing the strength of the knowledge gained from using a model developed to solve one problem to solve another using the same model. This study provides knowledge about how transfer learning of various pre-trained CNN models has been engaged in image classification by categorizing numerous cuisine dishes from Gujarat, India. It also evaluates the effectiveness of each pre-trained model. In this study, GoogleNet, VGG16, Resnet50, and EfficientNet models are compared. The dataset created for this study includes 10 Gujarati dishes.