<< ANN_train keras_toolbox image_model_test >>

keras_toolbox >> keras_toolbox > image_model

image_model

This function is used to develop a transfer-learning image recognition model.

Syntax

y = image_model(in1);

Parameters

y :

training accuracy of the model

in1 :

Number of epochs

Description

Upon calling this function, a dialog box will open which will ask you to choose a directory containing image files used to train the model. The image directory structure should by as follows.

example:

pets/

  dogs/

    dog001.jpg

    dog002.jpg

    ...

  cats/

    cat001.jpg

    cat002.jpg

where dogs and cats are the names of your image. Both class names and image names must be unique. So if such a directory structure is present, the folder "pets" must be selected. Supported image formats are jpg, jpeg and png. The folder 'alphabets' is provided with this zip, containing image dataset of A, B and C letters. Select this folder when asked for image directory by the dialog boox.

Examples

//Testing the neural network for input values sample to guess class name
y = image_model(10)

Authors


Report an issue
<< ANN_train keras_toolbox image_model_test >>