feat: filter for helping users with deuteranomaly and/or protanomaly (#1

)

* format code

* adjust filters

* feat: add color blind correction algorithm

README.md

@@ -2,29 +2,94 @@
 
 <img src="logo/logo.jpeg" width="200">
 
-The ByakuganVisualizer repository hosts a Python tool designed to compare images and highlight their differences. 
+The ByakuganVisualizer repository hosts a Python tool designed to compare images and highlight their differences.
 It simplifies the process of identifying disparities between images, making it ideal for tasks like testing and quality 
-assurance. Additionally, it offers options for customization, which can be helpful for color-blind users.
+assurance.
+Moreover, it offers a color filter that can be used to correct images for **color-blind users**.
 
 
+## Installation
+
+```bash
+pip install byakuganvisualizer
+```
+
 ## Usage 
 
 ```
-usage: byakugan_vision [-h] [--version] --diff DIFF [--filter {red,blue,green,yellow}] [--out_dir OUT_DIR]
+usage: byakugan_vision [-h] [--version] [--diff DIFF] [--filter {red,blue,green,yellow}] [--images IMAGES] [--deuteranomaly DEUTERANOMALY]
+                       [--protanomaly PROTANOMALY] [--out_dir OUT_DIR]
 
-ByakuganVisualizer: Tool for comparing images and highlighting differences.
+ByakuganVisualizer: Tool for correcting the color palett for color blind people and highlighting differences of images.
 
 options:
   -h, --help            show this help message and exit
   --version             show program's version number and exit
-  --diff DIFF           String containing a list of tuples "Path_To_Image1a,Path_To_Image2a;Path_To_Image1b,Path_To_Image2b...". Each tuple contains two paths to images to be compared.
+  --diff DIFF           String containing a list of tuples "Path_To_Image1a,Path_To_Image2a;Path_To_Image1b,Path_To_Image2b...". Each tuple
+                        contains two paths to images to be compared.
   --filter {red,blue,green,yellow}
                         Filter type (red, blue, green, yellow)
+  --images IMAGES       List of image names to be manipulated by a filter. E.g.: A,B,C,D
+  --deuteranomaly DEUTERANOMALY
+                        Expresses your degree of deuteranomaly, which will be used to correct the image. Default is 1.
+  --protanomaly PROTANOMALY
+                        Expresses your degree of protanomaly, which will be used to correct the image. Default is 1.
   --out_dir OUT_DIR     Output directory for the difference images
+```
+
+## Image Correction for Color Blind People
+
+In the following examples the image is corrected for deuteranomaly and protanomaly. 
+
+**Note:** The float values for deuteranomaly and protanomaly are between 0 and 10. The default value is 1.
+The used algorithm is based on the following paper: https://arxiv.org/abs/1711.10662.
+
+The image used in the example is from the following source:
+https://www.anime2you.de/news/606180/naruto-feiert-20-anime-jubilaeum/
+
+<img src="tests/test_images/naruto.jpg" style="width: 200px">
+
+### Deuteranomaly Correction
 
+```bash
+byakugan_vision --images "tests/test_images/naruto.jpg" --deuteranomaly 2
+```
+
+<img src="tests/test_images/filtered/Filtered_naruto_deuteranomaly_2_protanomaly_0.jpg" style="width: 200px">
+
+### Protanomaly Correction
+
+```bash
+byakugan_vision --images "tests/test_images/naruto.jpg" --protanomaly 2
 ```
 
-## Example
+<img src="tests/test_images/filtered/Filtered_naruto_deuteranomaly_0_protanomaly_2.jpg" style="width: 200px">
+
+### Deuteranomaly and Protanomaly Correction
+
+```bash
+byakugan_vision --images "tests/test_images/naruto.jpg" --deuteranomaly 2 --protanomaly 2
+```
+
+<img src="tests/test_images/filtered/Filtered_naruto_deuteranomaly_2_protanomaly_2.jpg" style="width: 200px">
+
+
+```bash
+byakugan_vision --images "tests/test_images/naruto.jpg" --deuteranomaly 0.5 --protanomaly 0.5
+```
+
+<img src="tests/test_images/filtered/Filtered_naruto_deuteranomaly_0.5_protanomaly_0.5.jpg" style="width: 200px">
+
+### Filter an Image
+
+```bash
+byakugan_vision --images "tests/test_images/naruto.jpg" --filter red
+```
+
+<img src="tests/test_images/filtered/Filtered_naruto_red.jpg" style="width: 200px">
+
+
+## Differences between images
 
 The left image used in the example is from the following source:
 https://www.anime2you.de/news/606180/naruto-feiert-20-anime-jubilaeum/

src/byakuganvisualizer/ImageFilter.py

@@ -14,7 +14,7 @@ def apply_red_filter(image_data):
         :return:
         """
         red_filtered = np.zeros_like(image_data)
-        red_filtered[:, :, 0] = image_data[:, :, 0] * 2
+        red_filtered[:, :, 0] = image_data[:, :, 0]
         return red_filtered
 
     @staticmethod
@@ -24,7 +24,7 @@ def apply_green_filter(image_data):
         :param image_data: The image data to apply the filter to.
         """
         green_filtered = np.zeros_like(image_data)
-        green_filtered[:, :, 1] = image_data[:, :, 1] * 2
+        green_filtered[:, :, 1] = image_data[:, :, 1]
         return green_filtered
 
     @staticmethod
@@ -35,7 +35,7 @@ def apply_blue_filter(image_data):
         :return:
         """
         blue_filtered = np.zeros_like(image_data)
-        blue_filtered[:, :, 2] = image_data[:, :, 2] * 2
+        blue_filtered[:, :, 2] = image_data[:, :, 2]
         return blue_filtered
 
     @staticmethod
@@ -46,6 +46,30 @@ def apply_yellow_filter(image_data):
         :return:
         """
         yellow_filtered = np.zeros_like(image_data)
-        yellow_filtered[:, :, 0] = image_data[:, :, 0] * 2
-        yellow_filtered[:, :, 1] = image_data[:, :, 1] * 2
+        yellow_filtered[:, :, 0] = image_data[:, :, 0]
+        yellow_filtered[:, :, 1] = image_data[:, :, 1]
         return yellow_filtered
+
+    @staticmethod
+    def correction_for_colorblindness(image_array, degree_protanomaly, degree_deuteranomaly):
+        """
+        Apply a colorblindness correction to the image data.
+        :param image_array:
+        :param degree_protanomaly:
+        :param degree_deuteranomaly:
+        :return:
+        """
+        r = image_array[..., 0]
+        g = image_array[..., 1]
+        b = image_array[..., 2]
+
+        corrected = np.copy(image_array)
+        r_corrected = (1 - degree_deuteranomaly / 2) * r + (degree_deuteranomaly / 2) * g
+        g_corrected = (degree_protanomaly / 2) * r + (1 - degree_protanomaly / 2) * g
+        b_corrected = ((degree_protanomaly / 4) * r + (degree_deuteranomaly / 4) * g +
+                       (1 - (degree_deuteranomaly + degree_protanomaly) / 4) * b)
+
+        corrected[..., 0] = r_corrected
+        corrected[..., 1] = g_corrected
+        corrected[..., 2] = b_corrected
+        return corrected

src/byakuganvisualizer/main.py

@@ -41,7 +41,8 @@ def parse_args(args):
       :obj:`argparse.Namespace`: command line parameters namespace
     """
     parser = argparse.ArgumentParser(
-        description="ByakuganVisualizer: Tool for comparing images and highlighting differences."
+        description="ByakuganVisualizer: Tool for correcting the color palett for color blind people and highlighting "
+                    "differences of images."
     )
     parser.add_argument(
         "--version",
@@ -51,7 +52,6 @@ def parse_args(args):
     parser.add_argument(
         '--diff',
         type=parse_tuples,
-        required=True,
         help='String containing a list of tuples "Path_To_Image1a,Path_To_Image2a;Path_To_Image1b,Path_To_Image2b...". '
              'Each tuple contains two paths to images to be compared.'
     )
@@ -60,13 +60,29 @@ def parse_args(args):
         choices=['red', 'blue', 'green', 'yellow'],
         help='Filter type (red, blue, green, yellow)'
     )
+    parser.add_argument(
+        '--images',
+        type=str,
+        help='List of image names to be manipulated by a filter. E.g.: A,B,C,D'
+    )
+    parser.add_argument(
+        '--deuteranomaly',
+        type=float,
+        default=0,
+        help='Expresses your degree of deuteranomaly, which will be used to correct the image. Default is 1.'
+    )
+    parser.add_argument(
+        '--protanomaly',
+        type=float,
+        default=0,
+        help='Expresses your degree of protanomaly, which will be used to correct the image. Default is 1.'
+    )
     parser.add_argument(
         '--out_dir',
         type=str,
         default='.',
         help='Output directory for the difference images'
     )
-    # TODO: option for just filtering an image
     return parser.parse_args(args)
 
 
@@ -77,33 +93,61 @@ def main(args):
         os.makedirs(args.out_dir)
         print(f"Output Directory: '{args.out_dir}' created.")
 
-    for pair in args.diff:
-        pair_name = os.path.basename(pair[0]).split('.')[0] + '_' + os.path.basename(pair[1]).split('.')[0]
-
-        image1 = Image.open(pair[0])
-        image2 = Image.open(pair[1])
-
-        array1 = np.array(image1)
-        array2 = np.array(image2)
-
-        # Calculate the absolute difference between the two arrays
-        difference_array = np.abs(array1 - array2)
-
-        if args.filter == 'red':
-            difference_array = ImageFilter.apply_red_filter(difference_array)
-            pair_name += '_red'
-        if args.filter == 'blue':
-            difference_array = ImageFilter.apply_blue_filter(difference_array)
-            pair_name += '_blue'
-        if args.filter == 'green':
-            difference_array = ImageFilter.apply_green_filter(difference_array)
-            pair_name += '_green'
-        if args.filter == 'yellow':
-            difference_array = ImageFilter.apply_yellow_filter(difference_array)
-            pair_name += '_yellow'
-
-        difference_image = Image.fromarray(difference_array.astype('uint8'))
-        difference_image.save(f'{args.out_dir}/Diff_{pair_name}.jpg')
+    if args.diff:
+        for pair in args.diff:
+            pair_name = os.path.basename(pair[0]).split('.')[0] + '_' + os.path.basename(pair[1]).split('.')[0]
+
+            image1 = Image.open(pair[0])
+            image2 = Image.open(pair[1])
+
+            array1 = np.array(image1)
+            array2 = np.array(image2)
+
+            # Calculate the absolute difference between the two arrays
+            difference_array = np.abs(array1 - array2)
+
+            if args.filter == 'red':
+                difference_array = ImageFilter.apply_red_filter(difference_array)
+                pair_name += '_red'
+            if args.filter == 'blue':
+                difference_array = ImageFilter.apply_blue_filter(difference_array)
+                pair_name += '_blue'
+            if args.filter == 'green':
+                difference_array = ImageFilter.apply_green_filter(difference_array)
+                pair_name += '_green'
+            if args.filter == 'yellow':
+                difference_array = ImageFilter.apply_yellow_filter(difference_array)
+                pair_name += '_yellow'
+
+            difference_image = Image.fromarray(difference_array.astype('uint8'))
+            difference_image.save(f'{args.out_dir}/Diff_{pair_name}.jpg')
+
+    if args.images:
+        args.images = args.images.split(',')
+        for img in args.images:
+            image_name = os.path.basename(img).split('.')[0]
+
+            image = Image.open(img)
+            rgb_array = np.array(image)
+
+            if args.filter == 'red':
+                rgb_array = ImageFilter.apply_red_filter(rgb_array)
+                image_name += '_red'
+            if args.filter == 'blue':
+                rgb_array = ImageFilter.apply_blue_filter(rgb_array)
+                image_name += '_blue'
+            if args.filter == 'green':
+                rgb_array = ImageFilter.apply_green_filter(rgb_array)
+                image_name += '_green'
+            if args.filter == 'yellow':
+                rgb_array = ImageFilter.apply_yellow_filter(rgb_array)
+                image_name += '_yellow'
+            if args.protanomaly > 0 or args.deuteranomaly > 0:
+                rgb_array = ImageFilter.correction_for_colorblindness(rgb_array, args.protanomaly, args.deuteranomaly)
+                image_name += f'_deuteranomaly_{args.deuteranomaly}_protanomaly_{args.protanomaly}'
+
+            filtered_image = Image.fromarray(rgb_array.astype('uint8'))
+            filtered_image.save(f'{args.out_dir}/Filtered_{image_name}.jpg')
 
 
 def run():