MnistImagePipelineLoadChooser.java

package org.deeplearning4j.examples.dataexamples;

import org.datavec.image.loader.NativeImageLoader;
import org.deeplearning4j.nn.multilayer.MultiLayerNetwork;
import org.deeplearning4j.util.ModelSerializer;
import org.nd4j.linalg.api.ndarray.INDArray;
import org.nd4j.linalg.dataset.api.preprocessor.DataNormalization;
import org.nd4j.linalg.dataset.api.preprocessor.ImagePreProcessingScaler;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

import javax.swing.*;
import java.io.File;
import java.util.Arrays;
import java.util.List;

/**
 * /**
 *  This code example is featured in this youtube video
 *
 *  http://www.youtube.com/watch?v=DRHIpeJpJDI
 *
 * This differs slightly from the Video Example,
 * The Video example had the data already downloaded
 * This example includes code that downloads the data
 *
 *  Data is downloaded from
 *
 *
 *  wget http://github.com/myleott/mnist_png/raw/master/mnist_png.tar.gz
 *  followed by tar xzvf mnist_png.tar.gz
 * The Data Directory mnist_png will have two child directories training and testing
 * The training and testing directories will have directories 0-9 with
 * 28 * 28 PNG images of handwritten images
 *
 *
 *
 *
 *
 *  This examples builds on the MnistImagePipelineExample
 *  by giving the user a file chooser to test an image of their choice
 *  against the Nueral Net, will the network think this cat is an 8 or a 1
 *  Seriously you can test anything, but obviously the network was trained on handwritten images
 *  0-9 white digit, black background, so it will work better with stuff closer to what it was
 *  designed for
 *
 */
public class MnistImagePipelineLoadChooser {
    private static Logger log = LoggerFactory.getLogger(MnistImagePipelineLoadChooser.class);


    /*
    Create a popup window to allow you to chose an image file to test against the
    trained Neural Network
    Chosen images will be automatically
    scaled to 28*28 grayscale
     */
    public static String fileChose(){
        JFileChooser fc = new JFileChooser();
        int ret = fc.showOpenDialog(null);
        if (ret == JFileChooser.APPROVE_OPTION)
        {
            File file = fc.getSelectedFile();
            String filename = file.getAbsolutePath();
            return filename;
        }
        else {
            return null;
        }
    }

    public static void main(String[] args) throws Exception{
        int height = 28;
        int width = 28;
        int channels = 1;

        // recordReader.getLabels()
        // In this version Labels are always in order
        // So this is no longer needed
        //List<Integer> labelList = Arrays.asList(2,3,7,1,6,4,0,5,8,9);
        List<Integer> labelList = Arrays.asList(0,1,2,3,4,5,6,7,8,9);

        // pop up file chooser
        String filechose = fileChose().toString();

        //LOAD NEURAL NETWORK

        // Where to save model
        File locationToSave = new File("trained_mnist_model.zip");
        // Check for presence of saved model
        if(locationToSave.exists()){
            System.out.println("\n######Saved Model Found######\n");
        }else{
            System.out.println("\n\n#######File not found!#######");
            System.out.println("This example depends on running ");
            System.out.println("MnistImagePipelineExampleSave");
            System.out.println("Run that Example First");
            System.out.println("#############################\n\n");


            System.exit(0);
        }

        MultiLayerNetwork model = ModelSerializer.restoreMultiLayerNetwork(locationToSave);

        log.info("*********TEST YOUR IMAGE AGAINST SAVED NETWORK********");

        // FileChose is a string we will need a file

        File file = new File(filechose);

        // Use NativeImageLoader to convert to numerical matrix

        NativeImageLoader loader = new NativeImageLoader(height, width, channels);

        // Get the image into an INDarray

        INDArray image = loader.asMatrix(file);

        // 0-255
        // 0-1
        DataNormalization scaler = new ImagePreProcessingScaler(0,1);
        scaler.transform(image);
        // Pass through to neural Net

        INDArray output = model.output(image);

        log.info("## The FILE CHOSEN WAS " + filechose);
        log.info("## The Neural Nets Pediction ##");
        log.info("## list of probabilities per label ##");
        //log.info("## List of Labels in Order## ");
        // In new versions labels are always in order
        log.info(output.toString());
        log.info(labelList.toString());

    }



}