tfm.vision.preprocess_ops.center_crop_image_v2

Center crop a square shape slice from the input image.

tfm.vision.preprocess_ops.center_crop_image_v2(
    image_bytes,
    image_shape,
    center_crop_fraction: float = CENTER_CROP_FRACTION
)

It crops a square shape slice from the image. The side of the actual crop is 224 / 256 = 0.875 of the short side of the original image. References: [1] Very Deep Convolutional Networks for Large-Scale Image Recognition https://arxiv.org/abs/1409.1556 [2] Deep Residual Learning for Image Recognition https://arxiv.org/abs/1512.03385

This is a faster version of center_crop_image which takes the original image bytes and image size as the inputs, and partially decode the JPEG bytes according to the center crop.

image_bytes a Tensor of type string representing the raw image bytes.
image_shape a Tensor specifying the shape of the raw image.
center_crop_fraction a float of ratio between the side of the cropped image and the short side of the original image

cropped_image a Tensor representing the center cropped image.