空间金字塔池化在深度卷积网络中的应用任务书

 2021-11-08 22:05:01

1. 毕业设计（论文）主要目标：

当前深度卷积神经网络都需要输入的图像尺寸固定。

这种人为的需要导致面对任意尺寸和比例的图像或子图像时降低识别的精度。

通过给网络配上一个叫做“空间金字塔池化”(spatial pyramid pooling)的池化策略以消除上述限制。

剩余内容已隐藏，您需要先支付后才能查看该篇文章全部内容！

2. 毕业设计（论文）主要内容：

利用现有的深卷积神经网络(cnn)需要固定大小(例如,224 x224)输入图像。其要求是“人工”，这可能会减少识别精度的图像或任意子图片大小。在这工作中我们给网络配上一个叫做“空间金字塔池化”(spatial pyramid pooling,)的池化策略以消除上述限制。这种新的网络结构称为SPP-net，可以生成一个固定长度的表示而不管图像的大小规模。金字塔池化对物体的形变十分鲁棒。由于诸多优点，SPP-net可以普遍帮助改进各类基于CNN的图像分类方法。在ImageNet2012数据集上，SPP-net将各种CNN架构的精度都大幅提升，尽管这些架构有着各自不同的设计。在PASCAL VOC 2007和Caltech101数据集上，SPP-net使用单一全图像表示在没有调优的情况下都达到了最好成绩。SPP-net在物体检测上也表现突出。使用SPP-net，只需要从整张图片计算一次特征图（feature map），然后对任意尺寸的区域（子图像）进行特征池化以产生一个固定尺寸的表示用于训练检测器。这个方法避免了反复计算卷积特征。

剩余内容已隐藏，您需要先支付后才能查看该篇文章全部内容！

3. 主要参考文献

  1. LeCun, Yann et al. “Backpropagation Applied to Handwritten Zip Code Recognition.” Neural Computation 1 (1989): 541-551.
  2. Lazebnik, Svetlana et al. “Beyond Bags of Features: Spatial Pyramid Matching for Recognizing Natural Scene Categories.” 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR06) 2 (2006): 2169-2178.
  3. Zitnick, C. Lawrence and Piotr Dollr. “Edge Boxes: Locating Object Proposals from Edges.” ECCV (2014).
  4. Krizhevsky, Alex et al. “ImageNet Classification with Deep Convolutional Neural Networks.” Commun. ACM 60 (2012): 84-90.
  5. Russakovsky, Olga et al. “ImageNet Large Scale Visual Recognition Challenge.” International Journal of Computer Vision 115 (2015): 211-252.
  6. Forsyth, David A.. “Object Detection with Discriminatively Trained Part-Based Models.” IEEE Computer 47 (2014): 6-7.
  7. Sermanet, Pierre et al. “OverFeat: Integrated Recognition, Localization and Detection using Convolutional Networks.” CoRR abs/1312.6229 (2014): n. pag.
  8. Girshick, Ross B. et al. “Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation.” 2014 IEEE Conference on Computer Vision and Pattern Recognition (2014): 580-587.
  9. Sande, Koen E. A. van de et al. “Segmentation as selective search for object recognition.” 2011 International Conference on Computer Vision (2011): 1879-1886.
  10. Howard, Andrew G.. “Some Improvements on Deep Convolutional Neural Network Based Image Classification.” CoRR abs/1312.5402 (2014): n. pag.
  11. Zou, Will Y. et al. “Generic Object Detection With Dense Neural Patterns and Regionlets.” CoRR abs/1404.4316 (2014): n. pag.
  12. Lowe, David G.. “Distinctive Image Features from Scale-Invariant Keypoints.” International Journal of Computer Vision 60 (2004): 91-110.
  13. Taigman, Yaniv et al. “DeepFace: Closing the Gap to Human-Level Performance in Face Verification.” 2014 IEEE Conference on Computer Vision and Pattern Recognition (2014): 1701-1708.
  14. Razavian, Ali Sharif et al. “CNN Features Off-the-Shelf: An Astounding Baseline for Recognition.” 2014 IEEE Conference on Computer Vision and Pattern Recognition Workshops (2014): 512-519.
  15. Jgou, Herv et al. “Aggregating Local Image Descriptors into Compact Codes.” IEEE Transactions on Pattern Analysis and Machine Intelligence 34 (2012): 1704-1716.
  16. Sivic, Josef and Andrew Zisserman. “Video Google: A Text Retrieval Approach to Object Matching in Videos.” ICCV (2003).
  17. Grauman, Kristen and Trevor Darrell. “The pyramid match kernel: discriminative classification with sets of image features.” Tenth IEEE International Conference on Computer Vision (ICCV05) Volume 1 2 (2005): 1458-1465 Vol. 2.
  18. Szegedy, Christian et al. “Going deeper with convolutions.” 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2015): 1-9.

剩余内容已隐藏，您需要先支付 10元 才能查看该篇文章全部内容！立即支付

以上是毕业论文任务书，课题毕业论文、开题报告、外文翻译、程序设计、图纸设计等资料可联系客服协助查找。

您可能感兴趣的文章