Add image processing functions and plot image grid
Browse files- src/utils.py +68 -1
src/utils.py
CHANGED
|
@@ -1,4 +1,8 @@
|
|
| 1 |
import os
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
import numpy as np
|
| 3 |
from PIL import Image, ImageOps
|
| 4 |
|
|
@@ -76,4 +80,67 @@ def track_files(folder_path, extensions=('.jpg', '.jpeg', '.png')):
|
|
| 76 |
if extension.lower() in extensions:
|
| 77 |
file_list.append(file_path)
|
| 78 |
|
| 79 |
-
return file_list
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import os
|
| 2 |
+
|
| 3 |
+
import cv2
|
| 4 |
+
import matplotlib.image as mpimg
|
| 5 |
+
import matplotlib.pyplot as plt
|
| 6 |
import numpy as np
|
| 7 |
from PIL import Image, ImageOps
|
| 8 |
|
|
|
|
| 80 |
if extension.lower() in extensions:
|
| 81 |
file_list.append(file_path)
|
| 82 |
|
| 83 |
+
return file_list
|
| 84 |
+
|
| 85 |
+
|
| 86 |
+
|
| 87 |
+
def crop_circle_roi(image_path):
|
| 88 |
+
"""
|
| 89 |
+
Crop the circular Region of Interest (ROI) from a fundus image.
|
| 90 |
+
|
| 91 |
+
Args:
|
| 92 |
+
- image_path (str): Path to the fundus image.
|
| 93 |
+
|
| 94 |
+
Returns:
|
| 95 |
+
- cropped_roi (numpy.ndarray): The cropped circular Region of Interest.
|
| 96 |
+
"""
|
| 97 |
+
# Read the image
|
| 98 |
+
image = cv2.imread(image_path, cv2.IMREAD_COLOR)
|
| 99 |
+
|
| 100 |
+
# Convert the image to grayscale
|
| 101 |
+
gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
|
| 102 |
+
|
| 103 |
+
# Apply thresholding to binarize the image
|
| 104 |
+
_, thresholded_image = cv2.threshold(gray_image, 50, 255, cv2.THRESH_BINARY)
|
| 105 |
+
|
| 106 |
+
# Find contours in the binary image
|
| 107 |
+
contours, _ = cv2.findContours(thresholded_image, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
|
| 108 |
+
|
| 109 |
+
# Assuming the largest contour corresponds to the ROI
|
| 110 |
+
contour = max(contours, key=cv2.contourArea)
|
| 111 |
+
|
| 112 |
+
# Get the bounding rectangle of the contour
|
| 113 |
+
x, y, w, h = cv2.boundingRect(contour)
|
| 114 |
+
|
| 115 |
+
# Crop the circular ROI using the bounding rectangle
|
| 116 |
+
cropped_roi = image[y:y+h, x:x+w]
|
| 117 |
+
|
| 118 |
+
return cropped_roi
|
| 119 |
+
|
| 120 |
+
def plot_image_grid(image_paths, roi_crop=False):
|
| 121 |
+
"""
|
| 122 |
+
Create a grid plot with a maximum of 16 images.
|
| 123 |
+
|
| 124 |
+
Args:
|
| 125 |
+
- image_paths (list): A list of image paths to be plotted.
|
| 126 |
+
|
| 127 |
+
Returns:
|
| 128 |
+
- None
|
| 129 |
+
"""
|
| 130 |
+
num_images = min(len(image_paths), 16)
|
| 131 |
+
num_rows = (num_images - 1) // 4 + 1
|
| 132 |
+
fig, axes = plt.subplots(num_rows, 4, figsize=(12, 3 * num_rows))
|
| 133 |
+
|
| 134 |
+
for i, ax in enumerate(axes.flat):
|
| 135 |
+
if i < num_images:
|
| 136 |
+
if roi_crop:
|
| 137 |
+
img = crop_and_pad_image(image_paths[i])
|
| 138 |
+
else:
|
| 139 |
+
img = mpimg.imread(image_paths[i])
|
| 140 |
+
ax.imshow(img)
|
| 141 |
+
ax.axis('off')
|
| 142 |
+
else:
|
| 143 |
+
ax.axis('off')
|
| 144 |
+
|
| 145 |
+
plt.tight_layout()
|
| 146 |
+
plt.show()
|