Scholarlink

With today in a fast paced environment, staying nutritiously active is vital, and tracking our food intake is not always easy. In this work, we present SMARTBITE, a deep learning-based application for automatic food identification and calorie estimate to support intelligent dietary monitoring. The system utilizes MobileNetV2, an advanced CNN pre-trained on the ImageNet dataset, to categories food photographs into five classifications: The machine readable datasets contain apple, banana, beetroot, bell pepper, and watermelon. Additional fine tuning is performed on the model for a specialized food dataset using data augmentation techniques to improve accuracy and resistance. The system that suggests, identifies food items and based on that suggested system provides real time calorie and nutrition data. Based on the evaluation of the test set, the model achieves 94% accuracy showing its ability in reliable food classification. Autonomously identifying food items from photos, evaluating the caloric content, and providing users with valuable nutrition information are the aims of SMARTBITE to enhance user diet health.

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

SMARTBITE: Advanced Food Recognition and Caloric Assessment for Dietary Management

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • G. Balaji,
  • M. Vignesh

摘要

With today in a fast paced environment, staying nutritiously active is vital, and tracking our food intake is not always easy. In this work, we present SMARTBITE, a deep learning-based application for automatic food identification and calorie estimate to support intelligent dietary monitoring. The system utilizes MobileNetV2, an advanced CNN pre-trained on the ImageNet dataset, to categories food photographs into five classifications: The machine readable datasets contain apple, banana, beetroot, bell pepper, and watermelon. Additional fine tuning is performed on the model for a specialized food dataset using data augmentation techniques to improve accuracy and resistance. The system that suggests, identifies food items and based on that suggested system provides real time calorie and nutrition data. Based on the evaluation of the test set, the model achieves 94% accuracy showing its ability in reliable food classification. Autonomously identifying food items from photos, evaluating the caloric content, and providing users with valuable nutrition information are the aims of SMARTBITE to enhance user diet health.