[1]齐 妍,孙 涵.基于判别性特征增强的小样本细粒度图像识别[J].计算机技术与发展,2024,34(01):44-51.[doi:10. 3969 / j. issn. 1673-629X. 2024. 01. 007]
　QI Yan,SUN Han.Few-shot Fine-grained Image Recognition Based on Discriminative Feature Enhancement[J].,2024,34(01):44-51.[doi:10. 3969 / j. issn. 1673-629X. 2024. 01. 007]
点击复制

基于判别性特征增强的小样本细粒度图像识别()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
34
期数:
2024年01期
页码:
44-51
栏目:
媒体计算
出版日期:
2024-01-10

文章信息/Info

Title:
Few-shot Fine-grained Image Recognition Based on Discriminative Feature Enhancement
文章编号:
1673-629X(2024)01-0044-08
作者:
齐 妍; 孙 涵
南京航空航天大学 计算机科学与技术学院 / 人工智能学院,江苏 南京 211106
Author(s):
QI Yan; SUN Han
School of Computer Science and Techonology / Artificial Itelligence,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China
关键词:
小样本细粒度图像识别; 深度学习; 特征增强; 注意力机制; 视觉一致性
Keywords:
few-shot fine-grained image recognition; deep learning; feature enhancement; attention mechanism; visual consistency
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2024. 01. 007
摘要:
小样本细粒度图像识别是深度学习领域中一个热门的研究课题,其基本任务是在学习有限数量样本的情况下识别出某一大类下的子类别的图像。 得益于卷积神经网络的快速发展,小样本细粒度图像识别在精度方面取得了显著的成果,但其性能仍受限于同一子类图像间的高方差以及不同分类任务中判别性特征的差异性。 针对上述问题,提出了一种基于判别性特征增强的小样本细粒度图像识别算法(DFENet) 。 DFENet 设计了对称注意力模块来增强类内视觉一致性学习,从而减少背景的影响,提高同类样本之间共享的特征表示的权重。 此外,DFENet 引入通道维度的判别性特征增强模块,利用支持集样本中同类样本内和不同类样本之间的通道关系进一步挖掘适合于当前任务的判别性特征,以提高识别准确率。 在三个经典的细粒度数据集 CUB-200-2011,Stanford Dogs,Stanford Cars 上进行了广泛的实验。 实验结果表明,该方法均取得了有竞争性的结果。
Abstract:
Few-shot fine-grained image recognition is a popular research topic in the field of deep learning. Its basic task is to identifyimages of subcategories under a super class while learning a limited number of samples. Thanks to the rapid development of convolutionalneural networks,the accuracy of few - shot fine - grained image recognition has achieved remarkable results,but its performance is stilllimited by the high variance among images of the same subclass and the variability of discriminative features in different classification　tasks. To address the above problems, we propose a few - shot fine - grained image recognition algorithm ( DFENet ) based ondiscriminative feature enhancement. DFENet is designed with a symmetric attention module to enhance intra - class visual consistencylearning,thus reducing the influence of background and increasing the weight of feature representations shared among similar samples. Inaddition,DFENet introduces a discriminative feature enhancement module of channel dimension,and further mines discriminative featuressuitable for the current task by exploiting the channel relationships within similar samples and between samples of different classes in thesupport set to improve the recognition accuracy. Extensive experiments are conducted on three classical fine-grained datasets CUB-200-2011,Stanford Dogs,and Stanford Cars. It is showed that the proposed method all achieves competitive results.

相似文献/References:

[1]陈强锐,谢世朋.基于深度学习的肺部肿瘤检测方法[J].计算机技术与发展,2018,28(04):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
　CHEN Qiang-rui,XIE Shi-peng.Lung Cancer Detection Method Based on Deep Learning[J].,2018,28(01):201.[doi:10.3969/ j. issn.1673-629X.2018.04.043]
[2]施泽浩,赵启军.基于全卷积网络的目标检测算法[J].计算机技术与发展,2018,28(05):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
　SHI Ze-hao,ZHAO Qi-jun.Object Detection Algorithm Based on Fully Convolutional Neural Network[J].,2018,28(01):55.[doi:10.3969/j.issn.1673－629X.2018.05.013]
[3]黄法秀,张世杰,吴志红,等.数据增广下的人脸识别研究[J].计算机技术与发展,2020,30(03):67.[doi:10. 3969 / j. issn. 1673-629X. 2020. 03. 013]
　HUANG Fa-xiu,ZHANG Shi-jie,WU Zhi-hong,et al.Research on Face Recognition Based on Data Augmentation[J].,2020,30(01):67.[doi:10. 3969 / j. issn. 1673-629X. 2020. 03. 013]
[4]陈浩翔,蔡建明,刘铿然,等. 手写数字深度特征学习与识别[J].计算机技术与发展,2016,26(07):19.
　CHEN Hao-xiang,CAI Jian-ming,LIU Keng-ran,et al. Deep Learning and Recognition of Handwritten Numeral Features[J].,2016,26(01):19.
[5]高翔,陈志,岳文静,等.基于视频场景深度学习的人物语义识别模型[J].计算机技术与发展,2018,28(06):53.[doi:10.3969/ j. issn.1673-629X.2018.06.012]
　GAO Xiang,CHEN Zhi,YUE Wen-jing,et al.Human Semantic Recognition Model Based on Video Scene Deep Learning[J].,2018,28(01):53.[doi:10.3969/ j. issn.1673-629X.2018.06.012]
[6]贺飞翔,赵启军. 基于深度学习的头部姿态估计[J].计算机技术与发展,2016,26(11):1.
　HE Fei-xiang,ZHAO Qi-jun. Head Pose Estimation Based on Deep Learning[J].,2016,26(01):1.
[7]徐 融,邱晓晖.一种改进的 YOLO V3 目标检测方法[J].计算机技术与发展,2020,30(07):30.[doi:10. 3969 / j. issn. 1673-629X. 2020. 07. 007]
　XU Rong,QIU Xiao-hui.An Improved YOLO V3 Object Detection[J].,2020,30(01):30.[doi:10. 3969 / j. issn. 1673-629X. 2020. 07. 007]
[8]曾志平[] [],萧海东[],张新鹏[]. 基于DBN的金融时序数据建模与决策[J].计算机技术与发展,2017,27(04):1.
　ZENG Zhi-ping[] [],XIAO Hai-dong[],ZHANG Xin-peng[]. Modeling and Decision-making of Financial Time Series Data with DBN[J].,2017,27(01):1.
[9]李全兵,文 钊*,田艳梅*,等.基于 WGAN 的音频关键词识别研究[J].计算机技术与发展,2021,31(08):26.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 005]
　LI Quan-bing,WEN Zhao *,TIAN Yan-mei *,et al.Research on Audio Keywords Recognition Based on WassersteinGenerative Adversarial Network[J].,2021,31(01):26.[doi:10. 3969 / j. issn. 1673-629X. 2021. 08. 005]
[10]李宏林. 分析式纹理合成技术及其在深度学习的应用[J].计算机技术与发展,2017,27(11):7.
　LI Hong-lin. Analyzed Texture-synthesis Techniques and Their Applications in Deep Learning[J].,2017,27(01):7.

更新日期/Last Update: 2024-01-10