A mobile, lightweight, poll-based food identification system
作者：
Highlights：
• We embed a complete food recognition system in a mobile device based on a multi-ranking classification with multiple features.
• So far, no one else has proposed a complete embedded system in a camera?equipped device for food recognition, to the best of our knowledge.
• Multi-ranking classification was able to cope with different indoor illuminations.
• Top-3 food classification achieved between 84% and 100% of success rate.
摘要
Highlights•We embed a complete food recognition system in a mobile device based on a multi-ranking classification with multiple features.•So far, no one else has proposed a complete embedded system in a camera?equipped device for food recognition, to the best of our knowledge.•Multi-ranking classification was able to cope with different indoor illuminations.•Top-3 food classification achieved between 84% and 100% of success rate.
论文关键词：Food identification,Multi-hypothesis segmentation,Multi-ranking classification,Mobile device
论文评审过程：Received 30 October 2012, Revised 4 December 2013, Accepted 11 December 2013, Available online 19 December 2013.
论文官网地址：https://doi.org/10.1016/j.patcog.2013.12.006
