tfm.vision.augment.translate_bbox

Equivalent of PIL Translate in X/Y dimension that shifts image and bbox.

tfm.vision.augment.translate_bbox(
    image, bboxes, pixels, replace, shift_horizontal
)

image 3D uint8 Tensor.
bboxes 2D Tensor that is a list of the bboxes in the image. Each bbox has 4 elements (min_y, min_x, max_y, max_x) of type float with values between [0, 1].
pixels An int. How many pixels to shift the image and bboxes
replace A one or three value 1D tensor to fill empty pixels.
shift_horizontal Boolean. If true then shift in X dimension else shift in Y dimension.

A tuple containing a 3D uint8 Tensor that will be the result of translating image by pixels. The second element of the tuple is bboxes, where now the coordinates will be shifted to reflect the shifted image.

ValueError if applied to video.