Scene classification via triplet networks
Scene classification via triplet networks
复制标题
通过三元组网络进行场景分类
DOI：
10.1109/jstars.2017.2761800
复制
发表时间：
2018
影响因子：
5.5
通讯作者：
Chao Huang
中科院分区：
文献类型：
--
作者：
Yishu Liu;Chao Huang
场景分类是遥感图像自动理解的一项基础工作。近年来，卷积神经网络成为遥感界的研究热点，在场景分类方面取得了很大的成绩。深度卷积网络主要以监督的方式进行训练，需要大量的标记训练样本。然而，明确标记的遥感数据通常是有限的。为了解决这个问题，在本文中，我们提出了一种新的场景分类方法，通过三元组网络，使用弱标记的图像作为网络输入。此外，我们还对现有的三重网络损失函数进行了理论研究，分析了它们在训练过程中处理“难”和/或“易”三重网络的不同机制。此外，四个新的损失函数，旨在更加重视“硬”三元组，以提高分类精度。已经进行了大量的实验，实验结果表明，三重网络加上我们提出的损失实现了最先进的性能在场景分类任务。
Scene classification is a fundamental task for automatic remote sensing image understanding. In recent years, convolutional neural networks have become a hot research topic in the remote sensing community, and have made great achievements in scene classification. Deep convolutional networks are primarily trained in a supervised way, requiring huge volumes of labeled training samples. However, clearly labeled remote sensing data are usually limited. To address this issue, in this paper, we propose a novel scene classification method via triplet networks, which use weakly labeled images as network inputs. Besides, we initiate a theoretical study on the three existing loss functions for triplet networks, analyzing their different underlying mechanisms for dealing with “hard” and/or “easy” triplets during training. Furthermore, four new loss functions are constructed, aiming at laying more stress on “hard” triplets to improve classification accuracy. Extensive experiments have been conducted, and the experimental results show that triplet networks coupled with our proposed losses achieve a state-of-the-art performance in scene classification tasks.