nntoolbox.vision.utils.utils module

nntoolbox.vision.utils.utils.compute_output_shape(inp_dim: int, padding: int, kernel_size: int, dilation: int, stride: int) → int

nntoolbox.vision.utils.utils.cv2_read_image(image_path, to_float: bool = False, flag: int = 1) → numpy.ndarray

Read an image using cv2 and convert to RGB

Parameters

• image_path –

• to_float – whether to convert image to float dats type:

• flag – indicate mode for cv2 read image

Returns

nntoolbox.vision.utils.utils.gram_matrix(x)

nntoolbox.vision.utils.utils.is_image(filename)

Check if filename has valid extension

Parameters

filename –

Returns

boolean indicating whether filename is a valid image filename

nntoolbox.vision.utils.utils.pil_to_tensor(pil, device=None)

nntoolbox.vision.utils.utils.tensor_to_np(tensor: torch.Tensor) → numpy.ndarray

Convert the tensor image to numpy format

nntoolbox.vision.utils.utils.tensor_to_pil(tensor)