Function Differences with tf.io.decode_image

View Source On Gitee

tf.io.decode_image

tf.io.decode_image(
    contents,
    channels=None,
    dtype=tf.dtypes.uint8,
    name=None,
    expand_animations=True
)

For more information, see tf.io.decode_image.

mindspore.dataset.vision.c_transforms.Decode

class mindspore.dataset.vision.c_transforms.Decode(
    rgb=True
)

For more information, see mindspore.dataset.vision.c_transforms.Decode.

Differences

TensorFlow: Decode the raw image bytes into an image with the specified number of channels and data type. It supports decoding GIF images.

MindSpore: Decode the raw image bytes into a RGB image.

Code Example

# The following implements Decode with MindSpore.
import numpy as np
import mindspore.dataset as ds

image = np.fromfile("/tmp/file.jpg", dtype=np.uint8)
result = ds.vision.c_transforms.Decode()(image)

# The following implements decode_image with TensorFlow.
import tensorflow as tf

raw = tf.io.read_file("/tmp/file.jpg")
result = tf.io.decode_image(raw, channels=3)