# CV-CNN

**Repository Path**: anruoran/CV-CNN

## Basic Information

- **Project Name**: CV-CNN
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-03-18
- **Last Updated**: 2020-12-19

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

Complex-Valued Convolutional Neural Network and Its Application in PoSAR Image Classification   

Requirements:   
Matlab   

Experimental steps:     
The test demo is on Flevoland dataset acquired by the AIRSAR sensor with L band.   
run ./Test Demo/TestDemo.m to evaluate the performance of the CV-CNN in PolSAR image classification    
    
The CV-CNN folder stores the implementation process of CV-CNN.    

Reference: This CV-CNN code is created based on the DeepLearnToolbox, which is a real CNN open source toolbox.     
https://github.com/rasmusbergpalm/DeepLearnToolbox    

Datatsets:
The datasets adopted in our paper "Complex-Valued Convolutional Neural Network and Its Application in PoSAR Image Classification" can be downloaded by the website "https://earth.esa.int/web/polsarpro/data-sources/sample-datasets".
Meanwhile, the corresponding ground truth has been uploaded in this project.

Remark: The ground truth of Flevoland-1989 containing 14 classes is Label_Flevoland_14cls.mat     
        The ground truth of Flevoland-1991 containing 15 classes is Label_Flevoland_15cls.mat     
        The ground truth of Oberpfaffenhofen, Germany is Label_Germany.mat
    
    
Updata: The input T matrix of the three datasets has been upload in the following Baiduyun link.

        Sampling data can be obtained at a certain sampling rate and divided into training set and validation set
        
        链接: https://pan.baidu.com/s/1GpA8f37hochOuPMWmZOkmQ 密码: 6qhk