###
:2019,28(9):58-64
←前一篇   |   后一篇→
本文二维码信息
码上扫一扫!
基于Android和卷积神经网络的鸟类识别系统
(1.中国科学院 计算机网络信息中心, 北京 100190;2.中国科学院大学, 北京 100049)
Bird Identification System Based on Android and Convolutional Neural Network
(1.Computer Network Information Center, Chinese Academy of Sciences, Beijing 100190, China;2.University of Chinese Academy of Sciences, Beijing 100049, China)
摘要
图/表
参考文献
相似文献
本文已被:浏览 114次   下载 83
投稿时间:2019-02-25    修订日期:2019-03-08
中文摘要: 随着深度学习的广泛应用和智能移动设备的普及,将深度学习的应用迁移到移动设备上已经成为一种新的趋势.本文设计了一种基于安卓平台和轻量级卷积神经网络的鸟类识别系统,该系统不依赖任何外部的计算资源和存储资源.本文提出以轻量级卷积神经网络作为基础模型的三种模型融合方法,分别是加权平均融合、双线型融合和多图片单模型融合.本文详细介绍了三种融合方式的结构和优缺点,并且给出了模型选择和超参数选择的一些方法.实验结果表明模型融合的方式相比单模型而言,识别精度有显著提高,可以更好的应用到安卓移动设备上.
Abstract:With the widespread use of deep learning and the popularity of smart mobile devices, it has become a new trend that migrates deep learning applications to mobile devices. This study designs a bird identification system based on Android platform and lightweight convolutional neural network. The system does not rely on any external computing and storage resources. This study also proposes three model stacking methods based on lightweight convolutional neural network as the basic model, which is weighted average, bilinear stacking, and multi-picture and single model stacking. In this paper, we introduce three stacking methods’ structure, advantages, and disadvantages in detail. And we also give some selection methods of hyperparameters through experiments. The experimental results show that the model stacking is much better than the single model, and the accuracy of the model has been significantly improved, which can be better applied to Android mobile devices.
文章编号:     中图分类号:    文献标志码:
基金项目:中国科学院科技服务网络计划(STS计划)(Y82E01);国家科技部国家科技基础条件平台项目(DKA2017-12-02-18)
引用文本:
肖理想,罗泽.基于Android和卷积神经网络的鸟类识别系统.计算机系统应用,2019,28(9):58-64
XIAO Li-Xiang,LUO Ze.Bird Identification System Based on Android and Convolutional Neural Network.COMPUTER SYSTEMS APPLICATIONS,2019,28(9):58-64

用微信扫一扫

用微信扫一扫