/LDCNN

MatConvNet implementation for low dimensional CNN “Learning Low Dimensional Convolutional Neural Networks for High-Resolution Remote Sensing Image Retrieval”

Primary LanguageMATLAB

MatConvNet implementation for low dimensional CNN (LDCNN)

LDCNN is a CNN architecture consisting of common convolutional layers and the mlpconv layer used in Network in Network (NIN). This work is for research purpose only!

How to use

  • Download MatConvNet (1.0-beta23 and 24 are tested) and add it to the root directory. Compile MatConvNet and call root directory LDCNN.
  • Download the AID dataset and add it to the data directory, i.e. LDCNN/data/AID
  • Download the pretrained VGGM model and add it to LDCNN/model
  • Run getImgSamples to randomly select images from AID dataset to construct training and test set and then add them to the AID directory
  • Run ldcnn_train

Citation

If you use this work please cite our work:

@article{zhou2017learning, title={Learning Low Dimensional Convolutional Neural Networks for High-Resolution Remote Sensing Image Retrieval}, author={Zhou, Weixun and Newsam, Shawn and Li, Congmin and Shao, Zhenfeng}, journal={Remote Sensing}, volume={9}, number={5}, pages={489}, year={2017}, publisher={Multidisciplinary Digital Publishing Institute} }