Package boofcv.deepboof

Class ImageClassifierVggCifar10

java.lang.Object

boofcv.deepboof.BaseImageClassifier

boofcv.deepboof.ImageClassifierVggCifar10

All Implemented Interfaces:

ImageClassifier<Planar<GrayF32>>, ImageModelBase<Planar<GrayF32>>

public class ImageClassifierVggCifar10
extends BaseImageClassifier

Image classification using VGG network trained in CIFAR 10 data. On the CIFAR 10 training set it get has 89.9% accuracy. This dataset contains images in 10 categories and 32x32 images.

See Also:

• szagoruyko/cifar.torch

Nested Class Summary

Nested classes/interfaces inherited from interface boofcv.abst.scene.ImageClassifier

ImageClassifier.Score

Field Summary

Fields inherited from class boofcv.deepboof.BaseImageClassifier

categories, categoryBest, categoryScores, imageRgb, imageSize, imageType, massage, network, tensorInput, tensorOutput

Constructor Summary

Constructors

Constructor	Description
ImageClassifierVggCifar10()

Method Summary

Modifier and Type	Method	Description
void	loadModel(File directory)	Expects there to be two files in the provided directory: YuvStatistics.txt model.net
protected Planar<GrayF32>	preprocess(Planar<GrayF32> image)	Massage the input image into a format recognized by the network

Methods inherited from class boofcv.deepboof.BaseImageClassifier

classify, getAllResults, getBestResult, getCategories, getImageRgb, getInputType, innerProcess

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

ImageClassifierVggCifar10

public ImageClassifierVggCifar10()

Method Details

loadModel

public void loadModel(File directory)
               throws IOException

Expects there to be two files in the provided directory:
YuvStatistics.txt
model.net

Parameters:

directory - Directory containing model files

Throws:

IOException - Throw if anything goes wrong while reading data

preprocess

protected Planar<GrayF32> preprocess(Planar<GrayF32> image)

Description copied from class: BaseImageClassifier

Massage the input image into a format recognized by the network

Overrides:

preprocess in class BaseImageClassifier