基于深度学习的细粒度图像检索任务书
2020-02-20 16:08:36
1. 毕业设计(论文)主要内容:
不同于通用的图像分析任务,细粒度图像分析的所属类别和粒度更为精细,它需要能在更细分的类别下对物体进行识别。细粒度图像检索是2015年首次提出的概念,其采用以图搜图的形式。例如,给定一张“绿头鸭”的图像,则需要将同为“绿头鸭”的图像从众多不同类鸟类图像中检索出来;类似地,需要将“劳斯莱斯幻影”从包括劳斯莱斯其他车型的不同品牌不同车型的众多图像中检索出来。细粒度图像检索的难点,一是图像粒度非常细微;二是对细粒度图像而言,哪怕是属于同一子类的图像本身也具有形态、姿势、颜色、背景等巨大差异。可以说,细粒度图像检索是图像检索领域和细粒度图像分析领域的一项具有新鲜生命力的研究课题。近年来随着深度学习的快速发展,深度卷积层所提取的高水平特征相比于传统的全连接层,对图像中的目标物体具有更高质量的定位能力。本课题借鉴卷积层的特征实现对图像目标的无监督定位,并提取高质量的检索向量,用于细粒图像检索。
1、图像分类和检索的基础知识;
2、浅层神经网络的学习;
2. 毕业设计(论文)主要任务及要求
1、查阅不少于15篇的相关资料,其中英文文献不少于3篇,完成开题报告。
2、完成不少于5000字的英文文献翻译工作。
3、收集相关的原始数据,并进行数据的预处理工作。
3. 毕业设计(论文)完成任务的计划与安排
1-3周:查阅文献,完成开题报告
4-6周:总体设计,完成论文综述
7-10周:设计算法,功能模块设计
11-13周:编码和测试
14-15周:写论文,提交初稿,给老师检查,修改定稿,答辩。
4. 主要参考文献
[1] N. Zhang, J. Donahue, R. Girshick, and T. Darrell.Part-based R-CNNs for fine-grained category detection. In European Conferenceon Computer Vision, Part I, LNCS 8689, pages 834–849, Zurich, Switzerland, Sept.2014. Springer, Switzerland.
[2] S. Branson, G. V. Horn, S. Belongie, and P. Perona. Bird speciescategorization using pose normalized deep convolutional nets. In BritishMachine Vision Conference, pages 1–14, Nottingham, England, Sept. 2014.
[3] S. Branson, O. Beijbom, and S. Belongie. Efficient large-scalestructured learning. In IEEE Conference on Computer Vision and PatternRecognition, In Proceedings of IEEE Conference on Computer Vision and PatternRecognition, pages 1806–1813, Portland, Oregon, Jun. 2013.
[4] X.-S. Wei, C.-W. Xie and J. Wu. Mask-CNN: Localizing Parts andSelecting Descriptors for Fine-Grained Image Recognition. arXiv:1605.06878,2016.
[5] T. Xiao, Y. Xu, K. Yang, J. Zhang, Y. Peng, and Z. Zhang. Theapplication of two-level attention models in deep convolutional neural networkfor fine-grained image classification. In Proceedings of IEEE Conference onComputer Vision and Pattern Recognition, pages 842–850, Boston, MA, Jun.2015.
[6] M. Simon and E. Rodner. Neural activation constellations:Unsupervised part model discovery with convolutional networks. In Proceedingsof IEEE International Conference on Computer Vision, pages 1143–1151, Sandiago,Chile, Dec. 2015.
[7] T.-Y. Lin, A. RoyChowdhury, and S. Maji. Bilinear CNN models forfine-grained visual recognition. In Proceedings of IEEE InternationalConference on Computer Vision, pages 1449–1457, Sandiago, Chile, Dec.2015.
[8] L. Xie, J. Wang, B. Zhang, and Q. Tian. Fine-grained image search.IEEE Transactions on Multimedia, vol. 17, no. 5, pp. 636–647, 2015.
[9] X.-S. Wei, J.-H. Luo, J. Wu, and Z.-H. Zhou. Selective ConvolutionalDescriptor Aggregation for Fine-Grained Image Retrieval. IEEE Transactions onImage Processing, 2017, in press.
[10] G. E. Hinton. Learning distributed representations of concepts. InAnnual Conference of the Cognitive Science Society, Amherst, MA, 1986, pp.1–12.