"""
Implementation of global motion estimators.
"""
import numpy as np
import os
import scipy.ndimage
import scipy.spatial
from ..utils import *
def _hausdorff_distance(E_1, E_2):
# binary structure
diamond = np.array([[0, 1, 0], [1, 1, 1], [0, 1, 0]])
# extract only 1-pixel border line of objects
E_1_per = E_1^scipy.ndimage.morphology.binary_erosion(E_1, structure=diamond)
E_2_per = E_2^scipy.ndimage.morphology.binary_erosion(E_2, structure=diamond)
A = scipy.ndimage.morphology.distance_transform_edt(~E_2_per)[E_1_per].max()
B = scipy.ndimage.morphology.distance_transform_edt(~E_1_per)[E_2_per].max()
return np.max((A, B))
[docs]def globalEdgeMotion(frame1, frame2, r=6, method='hamming'):
"""Global motion estimation using edge features
Given two frames, find a robust global translation vector
found using edge information.
Parameters
----------
frame1 : ndarray
first input frame, shape (1, M, N, C), (1, M, N), (M, N, C) or (M, N)
frame2 : ndarray
second input frame, shape (1, M, N, C), (1, M, N), (M, N, C) or (M, N)
r : int
Search radius for measuring correspondences.
method : string
"hamming" --> use Hamming distance when measuring edge correspondence distances. The distance used in the census transform. [#f1]_
"hausdorff" --> use Hausdorff distance when measuring edge correspondence distances. [#f2]_
Returns
----------
globalMotionVector : ndarray, shape (2,)
The motion to minimize edge distances by moving frame2 with respect to frame1.
References
----------
.. [#f1] Ramin Zabih and John Woodfill. Non-parametric local transforms for computing visual correspondence. Computer Vision-ECCV, 151-158, 1994.
.. [#f2] Kevin Mai, Ramin Zabih, and Justin Miller. Feature-based algorithms for detecting and classifying scene breaks. Cornell University, 1995.
"""
frame1 = vshape(frame1)
frame2 = vshape(frame2)
assert(frame1.shape == frame2.shape)
T, M, N, C = frame1.shape
assert C == 1, "called with frames having %d channels. Please supply only the luminance channel." % (C,)
# if type bool, then these are edge maps. No need to convert them
if frame1.dtype != np.bool:
E_1 = canny(frame1)
else:
E_1 = frame1
if frame2.dtype != np.bool:
E_2 = canny(frame2)
else:
E_2 = frame2
distances = []
displacements = []
for dx in range(-r, r+1, 1):
for dy in range(-r, r+1, 1):
cimage = np.roll(E_2, dx, axis=0)
cimage = np.roll(cimage, dy, axis=1)
# smallest distance between a point of points found in cimage
if method == 'hamming':
distance = scipy.spatial.distance.hamming(np.ravel(cimage), np.ravel(E_1))
elif method == 'hausdorff':
distance = _hausdorff_distance(cimage, E_2)
else:
raise Notimplemented
# compute # of bit flip distance
distances.append(distance)
displacements.append([dx, dy])
idx = np.argmin(distances)
return displacements[idx]
