用于飞行器检测与识别的卷积神经网络
计算机视觉与模式识别
2018-08-28 v1
摘要
本文研究使用文本引导的深度卷积神经网络分类器进行飞行器识别的问题。该网络接收一张航拍图像和一个期望类别,并通过匹配图像与期望类别的文本描述输出是或否。我们在合成航拍数据集上训练并测试我们的模型,期望类别由车辆的类型与颜色组合构成。当测试时考虑的类别多于训练时,该策略有所帮助。
引用
@article{arxiv.1808.08560,
title = {Convolutional Neural Networks for Aerial Vehicle Detection and Recognition},
author = {Amir Soleimani and Nasser M. Nasrabadi and Elias Griffith and Jason Ralph and Simon Maskell},
journal= {arXiv preprint arXiv:1808.08560},
year = {2018}
}
备注
This paper has been accepted in the National Aerospace Electronics Conference (NAECON) 2018 and would be indexed in IEEE