colorization package

Submodules

colorization.acl_model module

colorization.acl_resource module

colorization.colorize_process module

colorization.colorize_process.inference(model_path, input_image)

This function activate the model process after preprocess, and get result back.

model_path: str

the path of offline model(*.om file)

input_image: numpy array

resized image with L_channel obtained from preprocess.

resultnumpy array

ab channels of the preprocessed image.

colorization.colorize_process.postprocess(image_input_path, inference_result)
This function converts LAB image to BGR image (colorization).

It combines L channel obtained from source image and ab channels from Inference result.

Parameters:

image_input_pathstr

the path of the (gray) image to obtain L channel

inference_resultstr

Path to the .npy file containing the output of the inference function. (Consisting of ab channels)

return valuenumpy array

Colorized image.

colorization.colorize_process.preprocess(image_file)

This function downsizes the input image to modelWidth*modelHeight and extracts L_channel from the image.

input: image_file : numpy array

An image that is loaded from the specified file.

return valuenumpy array

L channel of the image subtracted by the 50 for mean-centering.

colorization.model_processor module

class colorization.model_processor.ModelProcessor(acl_resource, params)

Bases: object

__dict__ = mappingproxy({'__module__': 'colorization.model_processor', '__init__': <function ModelProcessor.__init__>, 'predict': <function ModelProcessor.predict>, '__dict__': <attribute '__dict__' of 'ModelProcessor' objects>, '__weakref__': <attribute '__weakref__' of 'ModelProcessor' objects>, '__doc__': None, '__annotations__': {}})
__init__(acl_resource, params)

Initialize self. See help(type(self)) for accurate signature.

__module__ = 'colorization.model_processor'
__weakref__

list of weak references to the object (if defined)

predict(model_input)

colorization.pipeline module

colorization.pipeline.colorize_image(image_input_path, image_output_path)

This function does the complete processing of a given image. It combines all of the subtasks together:

  • preprocess the image

  • colorize the image

  • postprocess the image

This function is called by the webservice.

image_input_pathstr

the path of the (gray) image to be processed

image_output_pathstr

the path of the (colorized) image after processing

return valueint

on success this function returns 0 on failure this function returns 1

colorize_image(‘/home/user/xyz/Pictures/pic1.jpg’,

‘/home/user/xyz/Pictures/pic1_colorized.jpg’)

The directories already exists, and the image can be directly written to the given output path.

colorization.pipeline.colorize_video(video_input_path, video_output_path)

This function does the complete processing of a given video. It combines all of the subtasks together:

  • split video into images

  • colorize each image

  • combine images to a video

  • if the original video has audio, add the audio to colorized video

This function is called by the webservice.

video_input_pathstr

the path of the (gray) video to be processed

video_output_pathstr

the path of the (colorized) video after processing

return valueint

on success this function returns 0 on failure this function returns 1

colorization.tests module

class colorization.tests.FunctionalTest(methodName='runTest')

Bases: unittest.case.TestCase

This class contains tests for the complete pipeline. (i.e. the function pipeline.colorize_image)

__module__ = 'colorization.tests'
setUp()

Hook method for setting up the test fixture before exercising it.

tearDown()

Hook method for deconstructing the test fixture after testing it.

test_colorize_twice()

Functional test to test the colorization of two images in single session

test_colorize_video()

Functional test to test the colorization of two images in single session

test_complete_colorize_image()

Functional test to test the complete colorizing process

class colorization.tests.PipelineTests(methodName='runTest')

Bases: unittest.case.TestCase

This class contains tests for each part of the pipeline See https://docs.python.org/3/library/unittest.html for more information.

__module__ = 'colorization.tests'
setUp()

Hook method for setting up the test fixture before exercising it.

tearDown()

Hook method for deconstructing the test fixture after testing it.

test_step_colorize_image()

Unit-Test to test the colorizing of an image

test_step_postprocess_image()

Unit-Test to test the postprocessing of an image

test_step_preprocess_image()

Unit-Test to test the preprocessing of an image

class colorization.tests.SplitAndMergeTestsForVideo(methodName='runTest')

Bases: unittest.case.TestCase

This class contains tests for the split and merge tests of video

__module__ = 'colorization.tests'
setUp()

Hook method for setting up the test fixture before exercising it.

tearDown()

Hook method for deconstructing the test fixture after testing it.

test_step_merge_audio_and_video()

Unit-Test to test the merge_audio_and_video function

test_step_split_audio_from_video()

Unit-Test to test the split_audio_from_video function

test_step_video2frames_frames2video()

Unit-Test to test the video2frames and frames2video function of a video

colorization.videodata module

colorization.videodata.frames2video(image_input_folder_path, video_output_path)

This function is used to convert images into a video. Args:

image_input_folder_path: path to the split images. video_output_path: path to the merged video

Returns:

0 for SUCCESS. 1 for FAILED.

colorization.videodata.merge_audio_and_video(video_input_path, audio_input_path, video_output_path)

This function is used to mge voice with a video merged from images. Args:

video_input_path: path of the origin video audio_input_path: path of the voice file video_output_path: path to the result video

Returns: int

on success this function returns 0 on failure this function returns 1

colorization.videodata.split_audio_from_video(video_input_path, audio_output_path)

This function is used to extract voice from a video. Args:

video_input_path: path of the origin video audio_output_path: path to the voice file

Returns: int

on success this function returns 0 on failure this function returns 1

colorization.videodata.video2frames(video_input_path, image_output_folder_path)

This function is used to convert video into images. Args:

video_input_path: filename of the video. image_output_folder_path: Output folder path containing images

Returns:

1 on fail. 0 on success.

Module contents