Automated Food image Classification using Deep Learning approach

Because of its growing importance in the health and medical fields, food image categorization is a new research subject. Automated food recognition techniques will undoubtedly aid in the development of diet monitoring systems, calorie estimation, and other similar applications in the future. Automated food classification methods based on deep learning algorithms are discussed in this research. For food image classification, SqueezeNet and VGG-16 Convolutional Neural Networks are utilized. It was shown that applying data augmentation and fine-tuning the hyper parameters improved the performance of these networks, making them appropriate for practical applications in the health and medical domains. Because SqueezeNet is a lightweight network, it is easy to set up and maintain. SqueezeNet can attain a high level of accuracy even with fewer parameters. Extraction of complex features from food photographs improves the accuracy of food image classification. The suggested VGG-16 network improves the performance of automatic food image classification. The planned VGG-16 has improved significantly as a result of increased network depth.

Keywords: Food Classification, Image processing, Squeeze Net, VGG-16 Network, Transfer learning.