keras_rmac:在Keras中实施RMAC

  • G7_760221
    了解作者
  • 2.3MB
    文件大小
  • zip
    文件格式
  • 0
    收藏次数
  • VIP专享
    资源类型
  • 0
    下载次数
  • 2022-05-15 23:08
    上传日期
凯拉斯(Keras)RMAC 基于(Tolias等人2016)和(Gordo等人2016),为Keras重新实现了区域最大卷积激活(RMAC)特征提取器。 该模型的架构如下图所示: RoiPooling代码来自: : 先决条件 此代码需要Keras 2.0或更高版本。 (2.7) (2.1.2) (0.9.0) ->下载文件并将其保存在data/文件夹中 参考 Tolias,G.,Sicre,R.和Jégou,H.具有CNN激活的积分最大池的特殊对象检索。 ICLR 2016。 Gordo,A.,Almazán,J.,Revaud,J.和&Larlus,D。深度图像检索:学习图像搜索的全局表示。 ECCV 2016。 引文 该代码是Keras的RMAC的重新实现。 如果使用此代码,请引用使用重新实现的论文和原始RMAC论文: @article{garcia2018a
keras_rmac-master.zip
  • keras_rmac-master
  • data
  • model.png
    122.4KB
  • sample.jpg
    238.5KB
  • PCAmatrices.mat
    1.9MB
  • rmac.py
    2.6KB
  • vgg16.py
    2.6KB
  • Readme.md
    1.8KB
  • get_regions.py
    1.5KB
  • RoiPooling.py
    4.8KB
  • utils.py
    629B
内容介绍
# Keras RMAC Re-implementation of Regional Maximum Activations of Convolutions (RMAC) feature extractor for Keras, based on (Tolias et al. 2016) and (Gordo et al. 2016). The architecture of the model is as in the image below: ![rmac](https://github.com/noagarcia/keras_rmac/blob/master/data/model.png?raw=true) RoiPooling code from: https://github.com/yhenon/keras-spp ## Prerequisites This code requires Keras version 2.0 or greater. - [Python][1] (2.7) - [Keras][2] (2.1.2) - [Theano][3] (0.9.0) - [VGG16 weights][4] --> download the file and save it in `data/` folder ## References - Tolias, G., Sicre, R., & Jégou, H. Particular object retrieval with integral max-pooling of CNN activations. ICLR 2016. - Gordo, A., Almazán, J., Revaud, J., & Larlus, D. Deep image retrieval: Learning global representations for image search. ECCV 2016. ## Citation This code is a re-implementation of RMAC for Keras. If using this code, please cite the paper where the re-implementation is used and the original RMAC paper: ``` @article{garcia2018asymmetric, author = {Noa Garcia and George Vogiatzis}, title = {Asymmetric Spatio-Temporal Embeddings for Large-Scale Image-to-Video Retrieval}, booktitle = {Proceedings of the British Machine Vision Conference}, year = {2018}, } ``` ``` @article{tolias2016particular, author = {Tolias, Giorgos and Sicre, Ronan and J{\'e}gou, Herv{\'e}}, title = {Particular object retrieval with integral max-pooling of CNN activations}, booktitle = {Proceedings of the International Conference on Learning Representations}, year = {2016}, } ``` [1]: https://www.python.org/download/releases/2.7/ [2]: https://keras.io/ [3]: http://deeplearning.net/software/theano_versions/0.9.X/ [4]: https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_th_dim_ordering_th_kernels.h5
评论
    相关推荐