Added filter parameters to readme.

README.md

@@ -76,20 +76,7 @@ Once all the requirements are installed, the program is ready to use. There are
 
 There is an option to input the filter series as a preset from json configuration file.
 
-
+<table style="width:100%;">
-<style>
-  table {
-    width: 100%;
-  }
-  th, td {
-    padding: 10px;
-    font-family: monospace;
-    width: 50%;
-    vertical-align: top;
-  }
-</style>
-
-<table>
   <thead>
     <tr>
       <th>General format</th>
@@ -147,49 +134,64 @@ All the filters used and their parameters are described below.
 ## Available filters with parameters
 
 -median blur
-    -ksize - kernel size (int)
+    -ksize - Kernel size (int)
 
 -gaussian blur
-    -ksize - Gaussian kernel size (int)
+    -sigma - Gaussian kernel standart deviation (int)
-    -sigmaX - Kernel deviation in X direction (float)
-    -sigmaY - Kernel deviation in Y direction (float) 
 
 -bilateral blur
-    -d - ? (int)
+    -diameter - Diameter of pixel neighborhood used for filtering (int)
-    -sigmaColor - ? (int)
+    -sigmaColor - Standard deviation for grayvalue/color distance (int)
-    -sigmaSpace - ? (int)
+    -sigmaSpace - Standard deviation for range distance in pixels (int)
 
--denoise
+-bilateral_scikit
-    -h - ? (int)
+    -sigmaColor - Standard deviation for grayvalue/color distance (float)
-    -tWS - template window size (int)
+    -sigmaSpace - Standard deviation for range distance in pixels (float)
-    -sWs - search window size (int)
 
--denoise_bilateral
+-nlmeans (non-local means) 
-    -sigmaColor - ? (int)
+    -patch_size - Size of patches used for denoising (int)
-    -sigmaSpace - ? (int)
+    -patch_distance - Distance in pixels where to search for patches (int)
-    -iterations - ? (int)
+    -h - Cut-off distance, higher means more smoothed image (float)
 
--denoise_tv_chambolle
+-total_variation
-    -weight - ? (float)
+    -weight - Denoising weight. (float)
-    -iterations - ? (int)
 
--sharpen 
+-block_match 
-    -kernel - ? (numpy.matrix)
+    -sigma - ? (?)
-
--unsharp mask
-    -strength - ? (float)
-    -ksize - kernel size (int)
 
 -unsharp mask scikit
-    -radius - ? (int)
+    -radius - Radius of the gaussian filter (int)
-    -amount - ? (float)
+    -amount - Strength of the unsharp mask (float)
-    
+    
--morph
+-farid
-    -kernel - ? (numpy.matrix)
+
-    -iterations - ? (int)
+-meijering
-    -op - opencv MORPH operation (MORPH_OPEN, MORPH_CLOSE,
+
-        MORPH_DILATE, MORPH_ERODE)
+-sato
-    -anchor - ? (tuple)
+
+-hessian
+    -sigmas - ? (float)
+
+-invert
+
+-scale_values
+
+-binarize
+    -threshold - value to cut differentiate pixels (int)
+    -maxval - maximal value (int) ??
+    -type - ? (str)
+
+-binarize_otsu
+
+-add_margin
+    -margin - number of pixels to add to the sides of the image (int)
+    -color - color value of newly added pixels (int)
+
+-erode
+    -kernel - kernel shape (numpy.matrix)
+
+-dilate
+    -kernel - kernel shape (numpy.matrix)
 
 # Comparison
 

src/filters.py

@@ -8,9 +8,10 @@ import cv2 as cv
 from skimage import filters as skiflt
 from skimage import restoration as skirest
 from skimage import morphology as skimorph
-# from scipy import signal as sig
+from scipy import ndimage
 from PIL import Image, ImageFilter
 import bm3d
+import matplotlib.pyplot as plt
 
 
 class filter:
@@ -133,7 +134,7 @@ class nlmeans(filter):
         # Size of patches used for denoising
         patch_size = int(params["patch_size"]) if params["patch_size"] else 5
 
-        # Distance in pixels where to search patches
+        # Distance in pixels where to search for patches
         patch_distance = int(params["patch_distance"]
                              ) if params["patch_distance"] else 3
 
@@ -210,7 +211,7 @@ class unsharp_mask_scikit(filter):
               str(radius) + " amount: " + str(amount))
         self.img = skiflt.unsharp_mask(self.img, radius=radius,
                                        amount=amount, channel_axis=None)
-        self.img = np.uint8(self.img * 255.0)  # converting back to uintknapsack
+        self.img = np.uint8(self.img * 255.0)  # converting back to uint
 
 # ------------------- EDGE DETECTION FILTERS -------------------#
 
@@ -319,6 +320,16 @@ class binarize(filter):
         self.img = cv.threshold(self.img, threshold, maxval, type)[1]
 
 
+class binarize_otsu(filter):
+    ''' Otsu binarization filter from opencv.
+    '''
+    def init(self, img):
+        super().__init__(img)
+    
+    def apply(self, _):
+        self.img = cv.threshold(self.img, 0, 255, cv.THRESH_BINARY + cv.THRESH_OTSU)[1]
+
+
 class add_margin(filter):
     def init(self, img):
         super().__init__(img)

src/main.py

@@ -122,7 +122,7 @@ class app:
         # TODO: possibly too bloated, sending all possible params to each filter
         # TODO: remove unnecessary params
         possible_params = {"h", "searchWindowSize", "templateWindowSize",
-                           "ksize", "kernel",
+                           "ksize", "kernel", "angle",
                            "sigmaColor", "sigmaSpace", "diameter", "anchor", "iterations",
                            "op", "strength", "amount", "radius", "weight", "channelAxis",
                            "theta", "sigma", "lambd", "gamma", "psi", "shape", "percent",
@@ -610,8 +610,7 @@ class app:
         '''Map fingerprint to finger model.
         '''
 
-        # TODO: this might be done in a better way
+        # TODO: this might be done in a better way, comment
-        # instead of summing up the values, use their product - 0 ?
         z = np.array([])
         for x in range(self.width):
             z = np.append(z, np.sqrt(1 - (2*x/self.width - 1)**2)