package¶

Subpackages¶

timbr wood classification flask app

app.predict(fullpath=None)[source]¶

predicts the material type of provided image

Parameters:	fullpath (str) – fullpath of image file
Returns:	dict of results
Return type:	dict

app.save_histogram(source, destination, colorspace='rgb')[source]¶

saves rgb or hsv hostograms of source image to destination

Parameters:	source (str) – fullpath of source image destination (str) – fullpath of destination image opt (colorspace) – colorspace of histogram acceptable values: ‘rgb’, ‘hsv’ default: ‘rgb’
Returns:	None
Return type:	None

contains the TextureClassifier class used for predicting the material type of a supplied texture

class model.TextureClassifier(db_path, model_name)[source]¶

classification_report(info)[source]¶

process each image in info and compile a classification report for all of them

Parameters:	info (DataFrame) – info object which lists images to be processed
Returns:	classification report
Return type:	DataFrame

get_data(fullpath)[source]¶

processes image file

Parameters:	fullpath (str) – fullpath to image file
Returns:	data
Return type:	DataFrame

get_results(pred)[source]¶

converts predictions into results

Parameters:	pred (numpy.array) – output of self._model.predict(data)
Returns:	list – results
Return type:	of dicts

info¶

compiled information about data in db_path/images and db_path/descriptions.json

Returns:	info object
Return type:	DataFrame

predict(filepath)[source]¶

predict the material type of provided image file

Parameters:	fullpath (str) – fullpath to image file
Returns:	list – results
Return type:	of dicts

set_model(filename)[source]¶

sets TextureClassifier’s internal model to given .pkl model

Returns:	None
Return type:	None