Python Examples of keras.applications.InceptionV3

Source File: image_utils.py From spark-deep-learning with Apache License 2.0

6 votes

def executeKerasInceptionV3(image_df, uri_col="filePath"):
    """
    Apply Keras InceptionV3 Model on input DataFrame.
    :param image_df: Dataset. contains a column (uri_col) for where the image file lives.
    :param uri_col: str. name of the column indicating where each row's image file lives.
    :return: ({str => np.array[float]}, {str => (str, str, float)}).
      image file uri to prediction probability array,
      image file uri to top K predictions (class id, class description, probability).
    """
    K.set_learning_phase(0)
    model = InceptionV3(weights="imagenet")

    values = {}
    topK = {}
    for row in image_df.select(uri_col).collect():
        raw_uri = row[uri_col]
        image = loadAndPreprocessKerasInceptionV3(raw_uri)
        values[raw_uri] = model.predict(image)
        topK[raw_uri] = decode_predictions(values[raw_uri], top=5)[0]
    return values, topK

Source File: tf_image_test.py From spark-deep-learning with Apache License 2.0

6 votes

def test_load_image_vs_keras_RGB(self):
        g = tf.Graph()
        with g.as_default():
            image_arr = utils.imageInputPlaceholder()
            # keras expects array in RGB order, we get it from image schema in BGR => need to flip
            preprocessed = preprocess_input(image_arr)

        output_col = "transformed_image"
        transformer = TFImageTransformer(channelOrder='RGB', inputCol="image", outputCol=output_col, graph=g,
                                         inputTensor=image_arr, outputTensor=preprocessed.name,
                                         outputMode="vector")

        image_df = image_utils.getSampleImageDF()
        df = transformer.transform(image_df.limit(5))

        for row in df.collect():
            processed = np.array(row[output_col], dtype = np.float32)
            # compare to keras loading
            images = self._loadImageViaKeras(row["image"]['origin'])
            image = images[0]
            image.shape = (1, image.shape[0] * image.shape[1] * image.shape[2])
            keras_processed = image[0]
            np.testing.assert_array_almost_equal(keras_processed, processed, decimal = 6)

    # Test full pre-processing for InceptionV3 as an example of a simple computation graph

Source File: tf_image_test.py From spark-deep-learning with Apache License 2.0

6 votes

def test_image_output(self):
        output_col = "resized_image"
        preprocessed_df = self._preprocessingInceptionV3Transformed("image", output_col)
        self.assertDfHasCols(preprocessed_df, [output_col])
        for row in preprocessed_df.collect():
            original = row["image"]
            processed = row[output_col]
            errMsg = "nChannels must match: original {} v.s. processed {}"
            errMsg = errMsg.format(original.nChannels, processed.nChannels)
            self.assertEqual(original.nChannels, processed.nChannels, errMsg)
            self.assertEqual(processed.height, InceptionV3Constants.INPUT_SHAPE[0])
            self.assertEqual(processed.width, InceptionV3Constants.INPUT_SHAPE[1])

    # TODO: add tests for non-RGB8 images, at least RGB-float32.

    # Test InceptionV3 prediction as an example of applying a trained model.

Source File: backend.py From head-detection-using-yolo with MIT License

6 votes

def __init__(self, input_size):
        input_image = Input(shape=(input_size, input_size, 3))

        inception = InceptionV3(input_shape=(input_size,input_size,3), include_top=False)
        inception.load_weights(INCEPTION3_BACKEND_PATH)

        x = inception(input_image)

        self.feature_extractor = Model(input_image, x)

Source File: file_util.py From Mosaicer with MIT License

5 votes

def make_model(model, image_size):
    if model == "inceptionv3":
        base_model = InceptionV3(include_top=False, input_shape=image_size + (3,))
    elif model == "vgg16" or model is None:
        base_model = VGG16(include_top=False, input_shape=image_size + (3,))
    elif model == "mobilenet":
        base_model = MobileNet(include_top=False, input_shape=image_size + (3,))
    return base_model

Source File: ml_model.py From machine_feeling with MIT License

5 votes

def load_model(self, model_path, pretrained=False):
        self.pretrained = pretrained

        if not pretrained:   
            try:
                if os.path.isfile(model_path):
                    self.model = load_model(model_path)
                    print('[+] Model loading complete')

                else:
                    print('[-] Model loading incomplete, could not find model - {}'.format(model_path))

            except Exception as err:
                print('[-] Model loading unsuccessful, please check your model file:')
                print(err)
        else:
            from keras.applications import InceptionV3
            self.model = InceptionV3(weights='imagenet')

        # a "begin" marker to time how long it takes (in real time) to compile
        start_compile = d.now()

        # actually compile the model
        self.model.compile(
            loss=l_type,
            optimizer=opt,
            metrics=met
        )

        # a calculation of the compile time, in seconds
        compile_time = (d.now() - start_compile).total_seconds()

        print('[+] Model successfully compiled in {:.3f} seconds'.format(compile_time))


    # a method for loading in the data given path (and many optional arguments)
    # note, this data path should point to a folder with the data