Skip to content

Make compatible for 1 channel only & Update frames handling #2

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
jonasfbraun wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Make compatible for 1 channel only & Update frames handling #2

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
8c306a6
included some 'if stack2 is not None else None' to make it compatible…
jonasfbraun Oct 8, 2020
00419ea
cleaned up how frames and ref_frame are handled in motion_compensate.…
jonasfbraun Apr 23, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
43 changes: 28 additions & 15 deletions ofco/main.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,7 @@

import os
import multiprocessing as mp
from tqdm import tqdm
from timeit import default_timer as timer
import numpy as np
from skimage import io
Expand Down Expand Up @@ -62,11 +63,8 @@ def apply_motion_field(stack1, stack2, w, frames):

stack1_warped = stack1

if stack2 is not None:
stack2_warped = stack2
else:
stack2_warped = None

stack2_warped = stack2
# TODO: Fine for now, but this only works if frames is a range from 0 to N_frames and the 0th frame is the reference frame.
for t in range(len(frames) - 1):
stack1_warped[t + 1, :, :] = bilinear_interpolate(
stack1[t + 1, :, :], w[t, :, :, 0], w[t, :, :, 1]
Expand Down Expand Up @@ -96,11 +94,13 @@ def motion_compensate(

start = timer()
stack1 = stack1[frames, :, :]
stack2 = stack2[frames, :, :]
if stack2 is not None:
stack2 = stack2[frames, :, :]
if ref_frame is not None:
stack1 = np.concatenate((ref_frame[np.newaxis], stack1), axis=0)
stack2 = np.concatenate((ref_frame[np.newaxis], stack2), axis=0)
frames = (0,) + tuple(frames)
if stack2 is not None:
stack2 = np.concatenate((ref_frame[np.newaxis], stack2), axis=0)
frames = tuple(np.arange(stack1.shape[0]))
stack1_rescale = (
(stack1 - np.amin(stack1)) / (np.amax(stack1) - np.amin(stack1)) * 255
)
Expand All @@ -117,7 +117,7 @@ def motion_compensate(
2,
)

i1 = stack1_rescale[frames[0], :, :]
i1 = stack1_rescale[0, :, :]
if parallel:
global global_stack1_rescale
global_stack1_rescale = stack1_rescale
Expand All @@ -127,15 +127,28 @@ def motion_compensate(
global_param = param
global global_initial_w
global_initial_w = initial_w
with mp.Pool(2) as pool:
w = pool.map(parallel_compute_motion, range(len(frames) - 1))
w = np.array(w)

N_process = 2
w = np.zeros(w_shape)
if verbose:
print("starting pool processing to compute motion fields of {} frames with {} processes".format(len(frames) - 1, N_process))
with mp.Pool(N_process) as pool:
if verbose:
for i, x_i in enumerate(tqdm(pool.imap(parallel_compute_motion,
range(len(frames) - 1)),
total=len(frames) - 1)):
w[i, :, :, :] = x_i
else:
w = pool.map(parallel_compute_motion, range(len(frames) - 1))
w = np.array(w)
del global_stack1_rescale
del global_i1
del global_param
else:
w = np.zeros(w_shape)
for t in range(len(frames) - 1):
if verbose:
print("starting non-parallel processing to compute motion fields")
for t in tqdm(range(len(frames) - 1)):
if verbose:
print("Frame {}\n".format(t))
i2 = stack1_rescale[t + 1, :, :]
Expand All @@ -160,9 +173,9 @@ def motion_compensate(
print("Time it took to warp images {}".format(end - start))

if ref_frame is not None:
w = w[1:]
stack1_warped = stack1_warped[1:]
stack2_warped = stack2_warped[1:]
if stack2_warped is not None:
stack2_warped = stack2_warped[1:]

if w_output is not None:
np.save(w_output, w)
Expand Down