• 设为首页
  • 点击收藏
  • 手机版
    手机扫一扫访问
    迪恩网络手机版
  • 关注官方公众号
    微信扫一扫关注
    公众号

filipradenovic/cnnimageretrieval: CNN Image Retrieval in MatConvNet: Training an ...

原作者: [db:作者] 来自: 网络 收藏 邀请

开源软件名称(OpenSource Name):

filipradenovic/cnnimageretrieval

开源软件地址(OpenSource Url):

https://github.com/filipradenovic/cnnimageretrieval

开源编程语言(OpenSource Language):

MATLAB 100.0%

开源软件介绍(OpenSource Introduction):

CNN Image Retrieval in MatConvNet: Training and evaluating CNNs for Image Retrieval in MatConvNet

This is a MATLAB toolbox that implements the training and testing of the approach described in our papers:

Deep Shape Matching,
Radenović F., Tolias G., Chum O., ECCV 2018 [arXiv]

Fine-tuning CNN Image Retrieval with No Human Annotation,
Radenović F., Tolias G., Chum O., TPAMI 2018 [arXiv]

CNN Image Retrieval Learns from BoW: Unsupervised Fine-Tuning with Hard Examples,
Radenović F., Tolias G., Chum O., ECCV 2016 [arXiv]

Prerequisites

In order to run this toolbox you will need:

  1. MATLAB (tested with MATLAB R2017a on Debian 8.1)
  2. MatConvNet MATLAB toolbox version 1.0-beta25
  3. All the rest (data + networks) is automatically downloaded with our scripts

Image retrieval

This code implements:

  1. Training (fine-tuning) CNN for image retrieval
  2. Learning supervised whitening for CNN image representations
  3. Testing CNN image retrieval on Oxford5k and Paris6k datasets

Run the following script in MATLAB:

>> run [MATCONVNET_ROOT]/matlab/vl_setupnn;
>> run [CNNIMAGERETRIEVAL_ROOT]/setup_cnnimageretrieval;
>> train_cnnimageretrieval;
>> test_cnnimageretrieval;

See [CNNIMAGERETRIEVAL_ROOT]/examples/train_cnnimageretrieval and [CNNIMAGERETRIEVAL_ROOT]/examples/test_cnnimageretrieval for additional details.

We provide the pretrained networks trained using the same parameters as in our ECCV 2016 and TPAMI 2018 papers. Performance comparison with the networks trained with our CNN Image Retrieval in PyTorch, on the original and the revisited Oxford and Paris benchmarks:

Model Oxford Paris ROxf (M) RPar (M) ROxf (H) RPar (H)
VGG16-GeM (MatConvNet) 87.9 87.7 61.9 69.3 33.7 44.3
VGG16-GeM (PyTorch) 87.2 87.8 60.5 69.3 32.4 44.3
ResNet101-GeM (MatConvNet) 87.8 92.7 64.7 77.2 38.5 56.3
ResNet101-GeM (PyTorch) 88.2 92.5 65.3 76.6 40.0 55.2

Note: Data and networks used for training and testing are automatically downloaded when using the example scripts.

Note (June 2022): We updated download files for Oxford 5k and Paris 6k images to use images with blurred faces as suggested by the original dataset owners. Bear in mind, "experiments have shown that one can use the face-blurred version for benchmarking image retrieval with negligible loss of accuracy".

Sketch-based image retrieval and shape matching

This code implements:

  1. Training (fine-tuning) CNN for sketch-based image retrieval and shape matching
  2. Testing CNN sketch-based image retrieval on Flickr15k dataset

Run the following script in MATLAB:

>> run [MATCONVNET_ROOT]/matlab/vl_setupnn;
>> run [CNNIMAGERETRIEVAL_ROOT]/setup_cnnimageretrieval;
>> train_cnnsketch2imageretrieval;
>> test_cnnsketch2imageretrieval;

See [CNNIMAGERETRIEVAL_ROOT]/examples/train_cnnsketch2imageretrieval and [CNNIMAGERETRIEVAL_ROOT]/examples/test_cnnsketch2imageretrieval for additional details.

We provide the pretrained networks trained using the same parameters as in our ECCV 2018 paper. The Flickr15k dataset used in the paper is slightly outdated compared to the latest one that is automatically downloaded when using this code (0.1 difference in mAP), so we report results here:

EdgeMAC components
Fine-tuned x x x x
Mirror x x
Multi-scale x x
mAP 42.0 43.5 45.7 46.2

Note: Data and networks used for testing are automatically downloaded when using the example scripts.

Related publications

Image retrieval

@article{RTC18a,
 title = {Fine-tuning {CNN} Image Retrieval with No Human Annotation},
 author = {Radenovi{\'c}, F. and Tolias, G. and Chum, O.}
 journal = {TPAMI},
 year = {2018}
}
@inproceedings{RTC16,
 title = {{CNN} Image Retrieval Learns from {BoW}: Unsupervised Fine-Tuning with Hard Examples},
 author = {Radenovi{\'c}, F. and Tolias, G. and Chum, O.},
 booktitle = {ECCV},
 year = {2016}
}

Sketch-based image retrieval and shape matching

@article{RTC18b,
 title = {Deep Shape Matching},
 author = {Radenovi{\'c}, F. and Tolias, G. and Chum, O.}
 journal = {ECCV},
 year = {2018}
}

Revisited benchmarks for Oxford and Paris ('roxford5k' and 'rparis6k')

@inproceedings{RITAC18,
 author = {Radenovi{\'c}, F. and Iscen, A. and Tolias, G. and Avrithis, Y. and Chum, O.},
 title = {Revisiting Oxford and Paris: Large-Scale Image Retrieval Benchmarking},
 booktitle = {CVPR},
 year = {2018}
}



鲜花

握手

雷人

路过

鸡蛋
该文章已有0人参与评论

请发表评论

全部评论

专题导读
热门推荐
阅读排行榜

扫描微信二维码

查看手机版网站

随时了解更新最新资讯

139-2527-9053

在线客服(服务时间 9:00~18:00)

在线QQ客服
地址:深圳市南山区西丽大学城创智工业园
电邮:jeky_zhao#qq.com
移动电话:139-2527-9053

Powered by 互联科技 X3.4© 2001-2213 极客世界.|Sitemap