基于深度卷积特征的细粒度图像分类研究综述-南京大学计算机系.PDFVIP

第XX 卷 第X 期 自 动 化 学 报 Vol. XX, No. X 201X 年X 月 ACTA AUTOMATICA SINICA Month, 201X 基于深度卷积特征的细粒度图像分类研究综述 1 1 罗建豪 吴建鑫 摘 要 细粒度图像分类问题是计算机视觉领域一项极具挑战的研究课题, 其目标是对子类进行识别, 如区分不同种类的鸟. 由于子类别间细微的类间差异和较大的类内差异, 传统的分类算法不得不依赖于大量的人工标注信息. 近年来, 随着深度学习 的发展, 深度卷积神经网络为细粒度图像分类带来了新的机遇. 大量基于深度卷积特征算法的提出, 促进了该领域的快速发展. 本文首先从该问题的定义以及研究意义出发, 介绍了细粒度图像分类算法的发展现状. 之后, 从强监督与弱监督两个角度对比 分析了不同算法之间的差异, 并比较了这些算法在常用数据集上的性能表现. 最后, 我们对这些算法进行了总结, 并讨论了该领 域未来可能的研究方向及其面临的挑战. 关键词 细粒度图像分类, 深度学习, 卷积神经网络, 计算机视觉 引用格式 罗建豪, 吴建鑫. 基于深度卷积特征的细粒度图像分类研究综述. 自动化学报, 201X, XX(X): X−X DOI 10.16383/j.aas.2017.c160425 A Survey on Fine-Grained Image Categorization Using Deep Convolutional Features LUO Jian-Hao1 WU Jian-Xin1 Abstract Fine-grained image categorization is a challenging task in the field of computer vision, which aims to classify sub-categories, such as different species of birds. Due to the low inter-class but high intra-class variation, traditional categorization algorithms have to depend on a large amount of annotation information. Recently, with the advances of deep learning, deep convolutional neural networks have opened up new opportunities for fine-grained recognition. Numerous deep convolutional feature-based algorithms have been proposed, which advance the development of fine-grained research. In this paper, starting from its definition, we give a brief introduction to some recent developments in fine-grained categorization. After that, we analyze different algorithms from the strongly supervised and weakly supervised aspects, and compare their performance on some popular datasets. Finally, we provide a brief summary of these methods as well as the potential future research directions and thei