Created
September 6, 2019 16:12
-
-
Save DoggettCK/8ba5bcee8c4d97aa67f09465c8f3307f to your computer and use it in GitHub Desktop.
OpenCV Python script to put the Matrix falling letters behind Taylor Swift for trueheart78
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
import os | |
import sys | |
import numpy as np | |
import cv2 | |
FRAME_DELAY = int(sys.argv[1]) if len(sys.argv) > 1 else 30 | |
foreground = cv2.VideoCapture('glowy_eyes.gif') | |
background = cv2.VideoCapture('matrix.gif') | |
fg_bg = cv2.createBackgroundSubtractorKNN() | |
def edge_detect(channel): | |
sobel_x = cv2.Sobel(channel, cv2.CV_16S, 1, 0) | |
sobel_y = cv2.Sobel(channel, cv2.CV_16S, 0, 1) | |
sobel = np.hypot(sobel_x, sobel_y) | |
sobel[sobel > 255] = 255; | |
return sobel | |
def find_significant_contours(frame, edge_image): | |
image, contours, hierarchy = cv2.findContours(edge_image, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) | |
# Find level 1 contours | |
level_1 = [] | |
for i, tupl in enumerate(hierarchy[0]): | |
# Each array is in format (next, prev, first child, parent) | |
# Filter ones without a parent | |
if tupl[3] == -1: | |
tupl = np.insert(tupl, 0, [i]) | |
level_1.append(tupl) | |
# From among them, find the contours with large surface area | |
significant = [] | |
too_small = edge_image.size * 0.025 | |
for tupl in level_1: | |
contour = contours[tupl[0]] | |
area = cv2.contourArea(contour) | |
if area > too_small: | |
significant.append([contour, area]) | |
# cv2.drawContours(frame, [contour], 0, (0, 255, 0), 1, cv2.LINE_AA, maxLevel=1) | |
significant.sort(key=lambda x: x[1]) | |
return [x[0] for x in significant] | |
def process_frame(fg_frame, bg_frame): | |
blurred = cv2.GaussianBlur(fg_frame, (11, 11), 0) | |
x, y = fg_frame.shape[:2] | |
# bg_frame = cv2.resize(bg_frame, (y, x)) | |
bg_frame = bg_frame[0:x, 0:y] | |
edge_image = np.max(np.array([ | |
edge_detect(blurred[:, :, 0]), | |
edge_detect(blurred[:, :, 1]), | |
edge_detect(blurred[:, :, 2]) | |
]), axis=0) | |
mean = 0.05 * np.mean(edge_image) | |
# Zero any value less than mean to reduce noise | |
edge_image[edge_image <= mean] = 0 | |
edge_image_8u = np.asarray(edge_image, np.uint8) | |
# Find contours | |
significant = find_significant_contours(fg_frame, edge_image_8u) | |
mask = edge_image.copy() | |
mask[mask > 0] = 0 | |
cv2.fillPoly(mask, significant, 255) | |
# invert mask | |
inv_mask = np.logical_not(mask) | |
# remove the background | |
fg_frame[inv_mask] = bg_frame[inv_mask] | |
return fg_frame | |
frames = [] | |
with open('image_list.txt', 'w') as file_list: | |
while(True): | |
fg_has_frame, fg_frame = foreground.read() | |
bg_has_frame, bg_frame = background.read() | |
if not all([fg_has_frame, bg_has_frame]): | |
break | |
processed = process_frame(fg_frame, bg_frame) | |
cv2.imshow('frame', processed) | |
filename = 'frame_{}.png'.format(len(frames)) | |
cv2.imwrite(filename, processed) | |
file_list.write('{}\n'.format(filename)) | |
frames.append(processed.copy()) | |
k = cv2.waitKey(FRAME_DELAY) & 0xff | |
if k == 27: | |
break | |
os.system('convert @image_list.txt glowy_output.gif') | |
foreground.release() | |
background.release() | |
cv2.destroyAllWindows() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment