Compartir a través de


microsoftml.extract_pixels: extrae píxeles de una imagen

Uso

microsoftml.extract_pixels(cols: [str, dict, list],
    use_alpha: bool = False, use_red: bool = True,
    use_green: bool = True, use_blue: bool = True,
    interleave_argb: bool = False, convert: bool = True,
    offset: float = None, scale: float = None, **kargs)

Descripción

Extrae los valores de píxeles de una imagen.

Detalles

extract_pixels extrae los valores de píxeles de una imagen. Las variables de entrada son imágenes del mismo tamaño, normalmente la salida de una transformación resizeImage. La salida son datos de píxeles en forma de vector que normalmente se usan como características para un aprendiz.

Argumentos

cols

Cadena de caracteres o lista de nombres de variables que se transformarán. Si es dict, las claves representan los nombres de las nuevas variables que se crearán.

use_alpha

Especifica si se debe usar el canal alfa. El valor predeterminado es False.

use_red

Especifica si se debe usar el canal rojo. El valor predeterminado es True.

use_green

Especifica si se debe usar el canal verde. El valor predeterminado es True.

use_blue

Especifica si se debe usar el canal azul. El valor predeterminado es True.

interleave_argb

Si se debe separar cada canal o intercalar en orden ARGB. Esto podría ser importante, por ejemplo, si está entrenando una red neuronal convolucional, ya que esto afectaría a la forma del kernel, el paso, etc.

convert

Si se debe convertir a número de punto flotante. El valor predeterminado es False.

offset

Especifica el desplazamiento (escalado previo). Esto requiere el valor convert = True. El valor predeterminado es None.

scale

Especifica el factor de escala. Esto requiere el valor convert = True. El valor predeterminado es None.

kargs

Argumentos adicionales que se envían al motor de proceso.

Devoluciones

Objeto que define la transformación.

Vea también

load_image, resize_image, featurize_image.

Ejemplo

'''
Example with images.
'''
import numpy
import pandas
from microsoftml import rx_neural_network, rx_predict, rx_fast_linear
from microsoftml import load_image, resize_image, extract_pixels
from microsoftml.datasets.image import get_RevolutionAnalyticslogo

train = pandas.DataFrame(data=dict(Path=[get_RevolutionAnalyticslogo()], Label=[True]))

# Loads the images from variable Path, resizes the images to 1x1 pixels
# and trains a neural net.
model1 = rx_neural_network("Label ~ Features", data=train, 
            ml_transforms=[            
                    load_image(cols=dict(Features="Path")), 
                    resize_image(cols="Features", width=1, height=1, resizing="Aniso"), 
                    extract_pixels(cols="Features")], 
            ml_transform_vars=["Path"], 
            num_hidden_nodes=1, num_iterations=1)

# Featurizes the images from variable Path using the default model, and trains a linear model on the result.
# If dnnModel == "AlexNet", the image has to be resized to 227x227.
model2 = rx_fast_linear("Label ~ Features ", data=train, 
            ml_transforms=[            
                    load_image(cols=dict(Features="Path")), 
                    resize_image(cols="Features", width=224, height=224), 
                    extract_pixels(cols="Features")], 
            ml_transform_vars=["Path"], max_iterations=1)

# We predict even if it does not make too much sense on this single image.
print("\nrx_neural_network")
prediction1 = rx_predict(model1, data=train)
print(prediction1)

print("\nrx_fast_linear")
prediction2 = rx_predict(model2, data=train)
print(prediction2)

Salida:

Automatically adding a MinMax normalization transform, use 'norm=Warn' or 'norm=No' to turn this behavior off.
Beginning processing data.
Rows Read: 1, Read Time: 0, Transform Time: 0
Beginning processing data.
Beginning processing data.
Rows Read: 1, Read Time: 0.001, Transform Time: 0
Beginning processing data.
Beginning processing data.
Rows Read: 1, Read Time: 0, Transform Time: 0
Beginning processing data.
Using: AVX Math

***** Net definition *****
  input Data [3];
  hidden H [1] sigmoid { // Depth 1
    from Data all;
  }
  output Result [1] sigmoid { // Depth 0
    from H all;
  }
***** End net definition *****
Input count: 3
Output count: 1
Output Function: Sigmoid
Loss Function: LogLoss
PreTrainer: NoPreTrainer
___________________________________________________________________
Starting training...
Learning rate: 0.001000
Momentum: 0.000000
InitWtsDiameter: 0.100000
___________________________________________________________________
Initializing 1 Hidden Layers, 6 Weights...
Estimated Pre-training MeanError = 0.707823
Iter:1/1, MeanErr=0.707823(0.00%), 0.00M WeightUpdates/sec
Done!
Estimated Post-training MeanError = 0.707499
___________________________________________________________________
Not training a calibrator because it is not needed.
Elapsed time: 00:00:00.2716496
Elapsed time: 00:00:00.0396484
Automatically adding a MinMax normalization transform, use 'norm=Warn' or 'norm=No' to turn this behavior off.
Beginning processing data.
Rows Read: 1, Read Time: 0, Transform Time: 0
Beginning processing data.
Beginning processing data.
Rows Read: 1, Read Time: 0, Transform Time: 0
Beginning processing data.
Beginning processing data.
Rows Read: 1, Read Time: 0, Transform Time: 0
Beginning processing data.
Using 2 threads to train.
Automatically choosing a check frequency of 2.
Auto-tuning parameters: L2 = 5.
Auto-tuning parameters: L1Threshold (L1/L2) = 1.
Using model from last iteration.
Not training a calibrator because it is not needed.
Elapsed time: 00:00:01.0508885
Elapsed time: 00:00:00.0133784

rx_neural_network
Beginning processing data.
Rows Read: 1, Read Time: 0, Transform Time: 0
Beginning processing data.
Elapsed time: 00:00:00.1339430
Finished writing 1 rows.
Writing completed.
  PredictedLabel     Score  Probability
0          False -0.028504     0.492875

rx_fast_linear
Beginning processing data.
Rows Read: 1, Read Time: 0, Transform Time: 0
Beginning processing data.
Elapsed time: 00:00:00.4977487
Finished writing 1 rows.
Writing completed.
  PredictedLabel  Score  Probability
0          False    0.0          0.5