Nowadays, the need for a healthy lifestyle has become increasingly essential due to the surge in unhealthy food options. Nutritional awareness is crucial in promoting overall well-being. A balanced diet includes consuming foods that provide essential nutrients and meet caloric requirements. Fruits are a vital source of micronutrients making them an important component of a healthy diet. Therefore, obtaining information about a fruit and its caloric content from a smart phone captured image can be highly beneficial. This capability can aid in dietary planning, help the individuals to make informed choices about their food intake, and maintain a balanced diet. In this research, we propose a fruit calorie measurement system that uses estimated volume from a single 2D image to calculate the amount of calories it contains. Our approach employs the YOLOv8 model for fruit detection and segmentation. For volume estimation, we take advantage of the almost symmetric shape of most of the fruits and use a modified disc integration method on each side of the fruit’s shape curvature. This volume is then used to calculate how much calories are contained in the fruit. We tested our system to estimate calories of the fruit "Apple". The experimental results show a reasonable accuracy for calorie estimation.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Novel Approach to Fruit Caloric Estimation Based on 2D Image-Driven Volume Measurement

  • Sayfullah Monsur Sujan Sarkar,
  • Kazi Fardin Islam,
  • Dhrubo Roy Partho,
  • Md. Golam Rashed,
  • Dipankar Das,
  • Vaskar Deka

摘要

Nowadays, the need for a healthy lifestyle has become increasingly essential due to the surge in unhealthy food options. Nutritional awareness is crucial in promoting overall well-being. A balanced diet includes consuming foods that provide essential nutrients and meet caloric requirements. Fruits are a vital source of micronutrients making them an important component of a healthy diet. Therefore, obtaining information about a fruit and its caloric content from a smart phone captured image can be highly beneficial. This capability can aid in dietary planning, help the individuals to make informed choices about their food intake, and maintain a balanced diet. In this research, we propose a fruit calorie measurement system that uses estimated volume from a single 2D image to calculate the amount of calories it contains. Our approach employs the YOLOv8 model for fruit detection and segmentation. For volume estimation, we take advantage of the almost symmetric shape of most of the fruits and use a modified disc integration method on each side of the fruit’s shape curvature. This volume is then used to calculate how much calories are contained in the fruit. We tested our system to estimate calories of the fruit "Apple". The experimental results show a reasonable accuracy for calorie estimation.