Scholarlink

As more people become mindful of healthy eating, fruits are becoming a staple in everyday meals. Mixed fruit plates are gaining popularity not just for their taste but also for their nutritional benefits. However, the types and amounts of fruits on a plate can significantly impact both calorie intake and overall nutrition. This chapter presents a practical method for analyzing images of fruit plates using a combination of deep learning and a random forest classifier. The goal is to automatically identify the types and quantities of fruits in an image and estimate their calorie content to support more informed dietary choices. To develop this system, a multi-label fruit recognition model was trained using the DeepFruit dataset which includes over 21,000 labeled images featuring 20 different fruit types in various combinations. The model is designed to perform well under real-world conditions such as varying lighting and angles. It also demonstrates high efficiency by processing each image in just 0.028 ms and achieves an accuracy of 96.34%. With these capabilities, the model shows strong potential for integration into health and nutrition applications that assist users in tracking their food intake and making healthier decisions.

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

Multi-label Fruit Recognition and Caloric Assessment for Healthy Eating Based Deep Learning

  • 章节

  • 发表日期：

    2026年

查看全文
  • Kai Xiao,
  • Rasha S. Aboul-Yazeed,
  • Ashraf Darwish,
  • Aboul Ella Hassanien

摘要

As more people become mindful of healthy eating, fruits are becoming a staple in everyday meals. Mixed fruit plates are gaining popularity not just for their taste but also for their nutritional benefits. However, the types and amounts of fruits on a plate can significantly impact both calorie intake and overall nutrition. This chapter presents a practical method for analyzing images of fruit plates using a combination of deep learning and a random forest classifier. The goal is to automatically identify the types and quantities of fruits in an image and estimate their calorie content to support more informed dietary choices. To develop this system, a multi-label fruit recognition model was trained using the DeepFruit dataset which includes over 21,000 labeled images featuring 20 different fruit types in various combinations. The model is designed to perform well under real-world conditions such as varying lighting and angles. It also demonstrates high efficiency by processing each image in just 0.028 ms and achieves an accuracy of 96.34%. With these capabilities, the model shows strong potential for integration into health and nutrition applications that assist users in tracking their food intake and making healthier decisions.