ASPCNet

We have relased the code of Adaptive Spatial Pattern Capsule Network (ASPCNet) algorithm, And the paper has been submitted to NEUROCOMPUTING.

Please cite us if our project is helpful to you

J. Wang, X. Tan, J. Lai, and J. Li. ASPCNet: Deep adaptive spatial pattern capsule network for hyperspectral image classification. Neurocomputing. 2022(486). 47-60.

If you have any questions, please contact us. e-mail: jinping_wang@foxmail.com

Requirements

 python3
 cuda=9
 cudnn=7
 tensorflow-gpu==1.12
 keras==2.2.4
 numpy 
 scipy

Validate_set is an optional parameter. If the value of Validate_set is set, the early_stop operation should be set as follows:

 callback = callbacks.EarlyStopping(monitor='val_acc',
                                    min_delta=0,
                                    patience=args.patience,
                                    verbose=1,
                                    mode='auto',
                                    restore_best_weights=True)

Otherwise, it should be judged based on the training accuracy or training loss, and be set as:

callback = callbacks.EarlyStopping(monitor='acc',
                                   min_delta=0,
                                   patience=args.patience,
                                   verbose=1,
                                   mode='auto',
                                   restore_best_weights=True)

If you want to run this code on the other datasets, please directly replace the *******.mat file in the data folder.

 --> data
    --> ****.mat
 --> img
    --> result
 --> ASP.py
 --> ASPCaps.py
 --> ASPCNet.py
 --> util.py

Quick start

 data source files and introduce can been found in http://www.ehu.eus/ccwintco/index.php?title=Hyperspectral_Remote_Sensing_Scenes

hookk/ASPCNet_HSIC

ASPCNet