'''
This module is used for resizing pictures containing molecules.
'''
import numpy as np
import cv2
import matplotlib.pyplot as plt
import os
from typing import Optional, Dict, Union
def _analyse_img(file, plot=False):
'''
Function used for analysing and getting useful information from an image.
This includes circle locations and sizes.
'''
# Read image
img = cv2.imread(file, cv2.IMREAD_UNCHANGED)
img = _remove_padding(img)
# Smooth it
img_copy = img.copy()
# Convert to greyscale
img_gray = cv2.medianBlur(img_copy, 3)
img_gray = cv2.cvtColor(img_gray, cv2.COLOR_BGRA2GRAY)
# Apply Hough transform to greyscale image
circles = cv2.HoughCircles(img_gray, cv2.HOUGH_GRADIENT_ALT, 1.5, 100,
param1=400, param2=0.8, minRadius=0, maxRadius=800)
circles = np.uint16(np.around(circles))[0]
circles = circles[(-circles[:, 2]).argsort()]
if plot:
img_copy = cv2.cvtColor(img_copy, cv2.COLOR_BGR2RGB)
for i, circle in enumerate(circles):
# draw the outer circle
cv2.circle(img_copy,(circle[0],circle[1]),circle[2],(0,255,0),2)
# draw the center of the circle
cv2.circle(img_copy,(circle[0],circle[1]),2,(0,0,255),3)
cv2.putText(img_copy, str(i), (circle[0] - 40, circle[1] + 44), cv2.FONT_HERSHEY_SIMPLEX, 4, (0, 0, 0), 16, cv2.LINE_AA)
cv2.putText(img_copy, str(i), (circle[0] - 40, circle[1] + 44), cv2.FONT_HERSHEY_SIMPLEX, 4, (255, 255, 255), 8, cv2.LINE_AA)
plt.imshow(img_copy)
plt.draw()
plt.pause(0.001)
return circles, img
def _remove_padding(img):
'''
Function used to remove padding from an image.
'''
gray = cv2.cvtColor(img, cv2.COLOR_BGRA2GRAY)
gray = 255*(gray < 240).astype(np.uint8) # To invert the text to white
coords = cv2.findNonZero(gray) # Find all non-zero points (text)
x, y, w, h = cv2.boundingRect(coords) # Find minimum spanning bounding box
rect = img[y:y+h, x:x+w] # Crop the image - note we do this on the original image
return rect
[docs]
def resize(img_paths, circle_numbers: Optional[Dict] = None, padding: Union[str, int, float] = 0):
'''
The main function for this module.
Takes a directory `d` and selected circles and resizes and moves images in order to produce new aligned images.
Args:
d: the directory to take images from.
circle_numbers: a dictionary containing the image names as keys and circle indices as the values.
The new images will be aligned based on these selected circles.
padding: the amount of padding to add to the new images.
If given an integer, it adds `padding` number of pixels.
Alternatively a string consisting of a number and a `%` sign can be given to add relative padding.
E.g. `padding='10%'` will add a 10% padding to the images.
'''
circles = {}
imgs = {}
for file in img_paths:
if file not in circle_numbers:
continue
circles_, img_ = _analyse_img(file)
circles[file] = circles_[circle_numbers[file]]
imgs[file] = img_
# get the atom sizes, which is simply the radius of the circle we selected
atom_sizes = {file: circle[2] for file, circle in circles.items()}
# calculate ratios for resizing the images
ratios = {file: max(atom_sizes.values())/size for file, size in atom_sizes.items()}
# resize old images to make new ones
imgs = {file: cv2.resize(img, [int(round(img.shape[1] * ratios[file])), int(round(img.shape[0] * ratios[file]))]) for file, img in imgs.items()}
# also resize circles to match the new ratio
circles = {file: circle * ratios[file] for file, circle in circles.items()}
# now find out where to place the circles
circle_x = [circle[0] for circle in circles.values()]
circle_y = [circle[1] for file, circle in circles.items()]
pad_left = {file: int((max(circle_x) - circle[0])) for file, circle in circles.items()}
pad_top = {file: int((max(circle_y) - circle[1])) for file, circle in circles.items()}
pad = {file: [(pad_top[file], 0), (pad_left[file], 0), (0, 0)] for file in imgs}
imgs = {file: np.pad(img, pad[file], constant_values=0) for file, img in imgs.items()}
imsize_x = [img.shape[1] for img in imgs.values()]
imsize_y = [img.shape[0] for img in imgs.values()]
pad_right = {file: int(max(imsize_x) - img.shape[1]) for file, img in imgs.items()}
pad_bottom = {file: int(max(imsize_y) - img.shape[0]) for file, img in imgs.items()}
pad = {file: [(0, pad_bottom[file]), (0, pad_right[file]), (0, 0)] for file in imgs}
imgs = {file: np.pad(img, pad[file], constant_values=0) for file, img in imgs.items()}
# add final padding
if isinstance(padding, str) and padding.endswith('%'):
padding = float(padding.removesuffix('%'))/100
padding = int(padding * list(imgs.values())[0].shape[1]), int(padding * list(imgs.values())[0].shape[0])
else:
padding = int(padding), int(padding)
pad = {file: [(padding[0], padding[0]), (padding[1], padding[1]), (0, 0)] for file in imgs}
imgs = {file: np.pad(img, pad[file], constant_values=0) for file, img in imgs.items()}
ret = {}
for file, img in imgs.items():
new_file = os.path.join(os.path.split(file)[0], 'resized_' + os.path.split(file)[1])
cv2.imwrite(new_file, img)
ret[file] = new_file
return ret
TCutility