How to using custom colormap with pixel calculation, gray scale value by opencv and make it workable on camera [closed]

Question

I tried to implement the mcolor colormap in the paper at page 8 to 10.

I developed the code to process a image file and it works ok. Now, I want to adopt this image processing method to the camera which is captured by opencv. I implemented it by python and opencv. It looks like that it's not workable as the computing limitation. Can anyone give me the direction to improve it? I expected to adopt my image processing algorithm to all frames of camera. Thank you. Below is my complete source code of python

import numpy as np import cv2 cap = cv2.VideoCapture(0) while True: ret, img = cap.read() img_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) img_new = np.ones([img.shape[0], img.shape[1], 3], np.uint8) mcolor = np.ones([img.shape[0], img.shape[1], 3], np.uint8) for i in range(img.shape[0]): for j in range(img.shape[1]): val = img_gray[i][j] % 36 if img_gray[i][j] <= 72: if val >= 1 and val <= 6: mcolor[i][j] = (0, 0, 35) elif val >= 7 and val <= 12: mcolor[i][j] = (i * 10, 0, 0) elif val >= 13 and val <= 18: mcolor[i][j] = (1, 0, 35) elif val >= 19 and val <= 24: mcolor[i][j] = (i * 20, 0, i * 20) elif val >= 25 and val <= 30: mcolor[i][j] = (i * 10, i * 5, i * 5) elif val >= 31 and val <= 36: mcolor[i][j] = (i * 10, 0, i * 4) elif img_gray[i][j] >= 73 and img_gray[i][j] <= 240: if val >= 1 and val <= 6: mcolor[i][j] = (0, 0, i * 6) elif val >= 7 and val <= 12: mcolor[i][j] = (0, i * 6, i * 6) elif val >= 13 and val <= 18: mcolor[i][j] = (0, i * 6, 0) elif val >= 19 and val <= 24: mcolor[i][j] = (i * 6, i * 6, 0) elif val >= 25 and val <= 30: mcolor[i][j] = (i * 6, 0, i * 6) elif val >= 31 and val <= 36: mcolor[i][j] = (i * 6, 0, 0) elif img_gray[i][j] > 241: if val >= 1 and val <= 6: mcolor[i][j] = (0, i * 10, 30) elif val >= 7 and val <= 12: mcolor[i][j] = (0, i * 10, 30) elif val >= 13 and val <= 18: mcolor[i][j] = (0, i * 10, 30) elif val >= 19 and val <= 24: mcolor[i][j] = (i * 10, 0, 0) elif val >= 25 and val <= 30: mcolor[i][j] = (i * 10, 0, 0) elif val >= 31 and val <= 36: mcolor[i][j] = (i * 10, 0, 0) for i in range(img.shape[0]): for j in range(img.shape[1]): b1, g1, r1 = img[i][j] r2, g2, b2 = mcolor[i][j] r, g, b = ( min(round((int(r1) + int(r2) + 1) / 1.2), 255), min(round((int(g1) + int(g2) + 1) / 1.2), 255), min(round((int(b1) + int(b2) + 1) / 1.2), 255), ) img_new[i][j] = (r, g, b) cv2.imshow("live", img_new) # cv2.imwrite("img_new.jpeg", img_new) # cv2.waitKey(0) if cv2.waitKey(1) == ord("q"): break cap.release() cv2.destroyAllWindows() 

Answer 1

Looking at just the code, I find it difficult to see its purpose.
In particular, I don't find colormap:

Document your code. In the code.

Define a function for the mapping presented.
Find or define a function to apply it to an image.

For non-negative values, img_gray[i][j] % 36 may take on values from 0 up, but never 36 up.
Prefer

 if lower < val <= mid: pass # handle lower range elif val <= upper: # 1 comparison less than mid < val <= upper pass # handle upper range 

Consider a table-driven approach for populating mcolor.
While following the paper in separating cases that are handled identically increases recognisability, it sure looks off in the code without a comment. (Looks another error in the paper, at that - where does that where i >= 240. come from?)

While next to all the numeric literals in the paper look magic to me,
avoid magic numbers like 36 and 255: give them names.

There are three occurrences of min(round((int(r1) + int(r2) + 1) / 1.2), 255): define a function. (isn't round((int(r1) + int(r2) + 1) / 2) equal to ceil((int(r1) + int(r2)) / 2)?)