计算机科学与探索 ›› 2019, Vol. 13 ›› Issue (1): 106-115.DOI: 10.3778/j.issn.1673-9418.1801022

• 人工智能与模式识别 • 上一篇    下一篇

深度学习驱动的智能冰箱

张卫山1+,吕  浩1,张元杰1,徐  亮2,赵德海3,周杰韩4   

  1. 1. 中国石油大学(华东) 计算机与通信工程学院,山东 青岛 266580
    2. 北京科技大学 计算机与通信工程学院,北京 100083
    3. 澳大利亚国立大学 计算机科学学院,澳大利亚 堪培拉 ACT 0200
    4. 奥卢大学 信息技术与电子工程学院,芬兰 奥卢 FI-90014
  • 出版日期:2019-01-01 发布日期:2019-01-09

Deep Learning Based Smart Refrigerator

ZHANG Weishan1+, LV Hao1, ZHANG Yuanjie1, XU Liang2, ZHAO Dehai3, ZHOU Jiehan4   

  1. 1. College of Computer and Communication Engineering, China University of Petroleum, Qingdao, Shandong 266580, China
    2. College of Computer and Communication Engineering, University of Science and Technology Beijing, Beijing 100083, China
    3. Research School of Computer Science, Australian National University, Canberra ACT 0200, Australia
    4. Faculty of Information Technology and Electrical Engineering, University of Oulu, Oulu FI-90014, Finland
  • Online:2019-01-01 Published:2019-01-09

摘要: 食品识别是智能冰箱的核心技术之一,但冰箱中食品的种类繁多并且摆放较为随意,相互遮挡的现象比较严重,这给冰箱中的食品识别带来了诸多挑战。为了提高冰箱内食物的识别效率,以识别冰箱中的果蔬为切入点,提出了一种基于智能冰箱的数据采集、数据处理和果蔬识别的整体架构,以及一种在冰箱环境下的基于深度学习的数据融合的果蔬识别方法。使用这种方法有效提高了在冰箱环境下果蔬识别的准确率。通过对采集的大量数据进行实验,证明了该方法具有良好的性能和识别准确度,能有效解决冰箱环境下果蔬识别问题。

关键词: 深度学习, 智能冰箱, 模型融合, 果蔬识别, 数据融合

Abstract: Food recognition is one of the core technologies for a smart refrigerator. As there are many kinds of food in a refrigerator that may be in a mass, it is a challenge for food recognition in the smart refrigerator. In order to improve the efficiency of food recognition, this paper proposes an architecture of data collection, data processing and image recognition for smart refrigerators. In addition, this paper proposes an approach of fruits and vegetables recognition, which uses multi-model data fusion based on deep learning techniques. This solution remarkably improves recognition accuracy of fruits and vegetables. Extensive evaluations using a large number of food images show that the proposed approach has good performance and high recognition accuracy that can meet industrial requirements on food recognition in refrigerators.

Key words: deep learning, smart refrigerator, multi-model fusion, fruits and vegetables recognition, data fusion