Abstract: Food Recognition is a computer vision application that has gained huge research interest. Food recognition and classification is a major task in managing health conditions and most importantly in assisting Type 1 diabetic patients in taking decisions on the right food to eat that will not worsen their health conditions. Diabetes has become a global health challenge threatening the well-being of millions of people across the world. Food recognition systems aid diabetic patients in monitoring what they eat, managing their chronic health conditions, and improving their quality of life. This paper presents an extensive review of food recognition and classification methods to aid diabetic patients and the food geographical regions of available datasets already studied. The review explores existing mobile and Desktop food recognition systems and diabetes self-care management applications. The analysis presented in this paper gives the following new insights: the most performing food recognition Methodologies that have been developed; the existing food datasets and the unexplored research areas. The findings in the literature reviewed show that Convolutional Neural Network (CNN) recognition techniques are widely applied in food recognition and classification systems compared to the Bag of Features (BoF) method. Also, the main challenge in this review is the functionalities of the available diabetes applications in the market and it was discovered that none offers recognition of Nigeria Foods to aid diabetes patients.
Keywords: Diabetes, food recognition, food classification, glycemic index, sugar level, deep learning, convolutional Neural Network
|
DOI:
10.17148/IJARCCE.2022.11905
