Tool and Images

EasyClassify is the deep learning classification library of Open eVision (EClassifier class).

Deep Learning Studio

To create an EasyClassify project in Deep Learning Studio:

1.

Start Deep Learning Studio.

2.

Create a new project and select EasyClassify in the New deep learning project dialog.

Tool properties

Capacity

●

The Capacity of the neural network (default: Normal) represents the quantity of information that it is capable of learning.

□

We recommend you to first train a Normal network on your dataset.

►

If the learning is working:

□

Try to use the Small or the Large network to get the performance (inference speed and accuracy) that best matches your application.

►

If the learning does not work it means either that:

□

Your dataset image resolution is too small (below 128 × 128): try to use the Extra Small network that specifically targets such cases.

□

Your dataset is complex and difficult: try to use the Extra Large network that specifically targets such cases.

●

In the API:

□

Use the string parameter EClassifier::ModelType to select the model capacity.

□

Use EClassifier::SetModelType to set the capacity of your tool.

□

Use EClassifier::GetAvailableModelTypes to list the available models for training.

Use pretrained model

●

A pretrained model should give faster convergence time and better accuracy for all the datasets except for extremely large datasets (> 10 000 images).

●

A pretrained model works by initializing the neural network with the weights learned on a large and complex dataset.

□

This allows to transfer the knowledge of some universal features (such as edges, corners...) learned on this dataset to your dataset.

□

If you do not use a pretrained model, the neural network is initialized randomly.

●

In the API, use EClassifier::SetUsePretrainedModel(false) to disable the user of pretrained models.

NOTE:

You cannot use the histogram equalization with pretrained weights as it would lead to poor training results.

Compute the heatmap

●

If this option is checked, the tool computes the heatmap along with each result.

□

The heatmap is available with EClassificationResult::GetHeatmap or with EClassificationResult::GetHeatmap.

►

Note that this option makes the computation of the results a bit slower but twice faster than computing the result and then computing the heatmap separately.

●

In the API, use EClassifier::SetComputeHeatmapWithResult to enable or disable this option.

●

If this option is disabled, use the method EClassifier::GetHeatmap to get the heatmap for an image.

Image format and normalization

Width, Height, Image type and Enable automatic image reformat

●

The input image format must have the width, height and number of channels corresponding to the input of the neural network.

●

By default, a classifier uses the image format of the first image inserted in the training dataset:

□

All other images are automatically reformatted (anisotropic rescaling and conversion between color and grayscale).

□

If EClassifier::SetEnableAutomaticImageReformat(false) is called, the classifier throws an exception when attempting to train or classify an image that does not have the correct image format.

It is recommended to use a width and heigh of at least 224 × 224, as the performances may start to deteriorate if the height or/and the width are smaller.

●

In Deep Learning Studio, you can set the input image format in the Classifier properties of the tools that you create in the Tools tab.

●

In the API, you can also set manually the input image format with the methods SetWidth, SetHeight and SetChannels (1 channel for grayscale images and 3 channels for color images).

Image resolution

The input image format must have a resolution of at least 128 x 128 for the normal and the large capacity or 64 x 64 for the small capacity and at most 1024 x 1024.

For the best processing speed, use the lowest resolution at which your "objects of interest" are still recognizable.

●

If your original images are smaller than the minimum resolution, upscale them to a resolution higher or equal to 128 x 128.

●

If your original images are larger than the maximum resolution, lower the resolution:

□

If the "objects of interest" are still recognizable, explicitly set the input image format of the classifier to this lower resolution.

□

If the "objects of interest" are not recognizable, divide your original images into sub-windows and use these sub-windows to train the classifier and make predictions. This presents the additional advantage of localizing the "object of interest" inside the original image.

Enable histogram equalization

The classifier can also apply an histogram equalization to every input image:

□

In Deep Learning Studio, activate it in the image format controls in the Image properties and augmentation tab.

□

In the API, use EClassifier::SetEnableHistogramEqualization(true) to activate it.

NOTE:

You cannot use the histogram equalization with pretrained weights as it would lead to poor training results.