Food Segmentation and Calorie Estimation of Coastal Food Delicacies

Abstract

With the growing interest in using artificial intelligence (AI) for nutritional analysis, complex models have been developed to calculate the calorie content of various foods. In this study, we employ Machine Learning (ML) techniques for estimating calories in seafood dishes. To enhance accuracy and resilience, the model combines two popular algorithms: linear regression and k-nearest neighbors (KNN). Seaside food, renowned for its rich and varied flavors, presents a unique challenge in calorie estimation due to the variety of ingredients and cooking techniques used. The proposed machine learning model addresses these challenges by identifying complex patterns in the dataset while considering the unique qualities of coastal cuisine. The KNN algorithm, by finding local patterns in the dataset, enhances the model's efficacy, making it adept at capturing the subtleties of regional variations in coastal food. Additionally, the linear regression model complements the KNN approach by highlighting more general patterns and connections among different components, cooking methods, and caloric content. The training and assessment dataset comprise an extensive compilation of seaside cuisine dishes, each labeled with precise calorie counts. The model is trained to generalize from this data, enabling it to predict the calorie content of previously unseen dishes accurately. Performance evaluation indicates that the combined KNN and linear regression model outperforms individual algorithms in terms of accuracy and generalization across a variety of coastal cuisines.