turicreate.SArray.pixel_array_to_image

SArray.pixel_array_to_image(width, height, channels, undefined_on_failure=True, allow_rounding=False)

Create a new SArray with all the values cast to turicreate.image.Image of uniform size.

Parameters:

width: int

The width of the new images.

height: int

The height of the new images.

channels: int.

Number of channels of the new images.

undefined_on_failure: bool , optional , default True

If True, return None type instead of Image type in failure instances. If False, raises error upon failure.

allow_rounding: bool, optional , default False

If True, rounds non-integer values when converting to Image type. If False, raises error upon rounding.

Returns:

out : SArray[turicreate.Image]

The SArray converted to the type ‘turicreate.Image’.

See also

astype, str_to_datetime, datetime_to_str

Examples

The MNIST data is scaled from 0 to 1, but our image type only loads integer pixel values from 0 to 255. If we just convert without scaling, all values below one would be cast to 0.

>>> mnist_array = turicreate.SArray('https://static.turi.com/datasets/mnist/mnist_vec_sarray')
>>> scaled_mnist_array = mnist_array * 255
>>> mnist_img_sarray = tc.SArray.pixel_array_to_image(scaled_mnist_array, 28, 28, 1, allow_rounding = True)