Modern agricultural revolution has enabled farmers to use Artificial Intelligence (AI)-based technologies resulting in improved crop yield, reduction in cost and optimization of the agricultural processes. Most of the food crops and grains are cultivated through farming and each food has its own characteristics, texture, color, shape, size, and granularity. Recognizing and identifying food items is necessary for calorie and nutrition estimation in order to maintain healthy eating habits and good health. Since the food items come in various textures, contents, and structure, it is tedious to distinguish food material. Texture is a prominent feature usually observed in all the agricultural food grains, and it is difficult to classify food material without recognizing its texture. It is evident that texture analysis and recognition is a significant and vital component in agriculture-based applications. The intention of the work is to prove the significance of the deep architecture InceptionResNetv2 model integrated with machine learning classifiers and to assess the performance of the classifiers in food classification system. We tested our proposed model with challenging RawFooT food texture dataset with 68 classes using10 fold cross validation. Extensive ablation study and analysis demonstrated an excellent accuracy of 100% with linear discriminant analysis and 99.8% with Quadratic SVM classifier for RawFooT dataset.
CITATION STYLE
Simon, P., & Uma, V. (2023). Integrating InceptionResNetv2 Model and Machine Learning Classifiers for Food Texture Classification. In Cognitive Science and Technology (pp. 531–539). Springer. https://doi.org/10.1007/978-981-19-8086-2_51
Mendeley helps you to discover research relevant for your work.