An Optimized Recurrent Neural Network for re-modernize food dining bowls and estimating food capacity from images

N. Veena,M. Prasad,S. Aruna Deepthi,B. Swaroopa Rani,Manjushree Nayak,S. Someshwar
DOI: https://doi.org/10.1016/j.entcom.2024.100664
IF: 2.072
2024-03-27
Entertainment Computing
Abstract:Maintaining health and preventing chronic diseases requires awareness of the energy and nutrient content of one's diet. Due to the rapid development of computing and Artificial intelligence (AI) technology, nutritional evaluation can be performed using smartphones or wearable food images.Thechallenges of this technology include estimating the visual estimation of the amount of food in a bowl. We describe an innovative approach for determining a bowl's dimensions by sticking a paper ruler throughout the bottom and edges yet photographing the outcome. We used a variety of colored liquids, bowls and crystallized foods to evaluate the precision of our technique for measuring food volume employing spherical bowls as boxes. Different food-related datasets can have other numbers of images, types of food dining bowls and estimated food capacity from the images represented. Image preprocessing using a wiener filter enhances the estimating of food dining bowl capacity from images. Following preprocessing, a threshold-based approach to image segmentation is used for the resulting data; finally, we used the Gradient Lion Swarm Optimized Discrete Recurrent Neural Network (GLSO-DRNN) model to present an AI-based multi-dish food identification approach. An excellent mAP = 0.96 was found between the results. The suggested model shows great promise as a means of improving dish reporting because of its high classification performance.The GLSO-DRNN performed better to proving its flexibility and efficient discrete parameter optimization. In addition to offering accurate food quantity estimates, it demonstrated significant potential for other uses in the field of smart dining technology. The results highlight an effective of this advanced computational method that is improving food dining bowls and interactions with components associated with food. In the future, we plan to improve the techniques' performance and incorporate our technology into a working Smartphone based on the cloud environment.
computer science, cybernetics, interdisciplinary applications, software engineering
What problem does this paper attempt to address?