Add A2b

A2/GUI.py

@@ -154,8 +154,8 @@ class GUI(object):
 if __name__ == '__main__':
     # Todo: Add smooth filter
     """
-    This program is for editing images with ORB keypoints.
-    Those keypoints are then been grouped with k-means.
+    This program implements the CCV.
+    Since the DFS is very costly please take images of: /images 
 
     Preparation:
         Make sure you have already installed all packages by running:

A2/Graph.py

@@ -9,6 +9,15 @@ from KEYS import KEYS
 class Graph(object):
 
     def __init__(self, row, col, g, bins, threshold) -> None:
+        """
+        This constructs the graph for the CCV.
+
+        :param row: Amount of rows
+        :param col: Amount if columns
+        :param g: Graph for recursion
+        :param bins: Amount of bins
+        :param threshold: What is the threshold for plotting
+        """
         self.ROW = row
         self.COL = col
         self.graph = g
@@ -16,10 +25,28 @@ class Graph(object):
         self.bins = bins
         self.threshold = threshold
 
-    def isSafe(self, i, j, visited, g):
+    def isSafe(self, i, j, visited, g) -> bool:
+        """
+        Check if the neighbor is in the matrix.
+
+        :param i: row coordinate
+        :param j: column coordinate
+        :param visited: matrix which holds the visited fileds
+        :param g: graph
+        :return: Is the neighbor in the matrix ?
+        """
         return 0 <= i < self.ROW and 0 <= j < self.COL and not visited[i][j] and g[i][j]
 
-    def DFS(self, i, j, visited, g):
+    def DFS(self, i, j, visited, g) -> None:
+        """
+        Depth first search to get the connected components in the matrix.
+
+        :param i: row coordinate
+        :param j: column coordinate
+        :param visited: matrix which holds the visited fileds
+        :param g: graph
+        :return: None
+        """
         rowNbr = [-1, -1, -1, 0, 0, 1, 1, 1]
         colNbr = [-1, 0, 1, -1, 1, -1, 0, 1]
 
@@ -29,7 +56,15 @@ class Graph(object):
             if self.isSafe(i + rowNbr[k], j + colNbr[k], visited, g):
                 self.DFS(i + rowNbr[k], j + colNbr[k], visited, g)
 
-    def label_connected_components(self, visited, label, count):
+    def label_connected_components(self, visited, label, count) -> None:
+        """
+        Label the connected components.
+
+        :param visited: matrix which holds the visited fileds
+        :param label: label of the color
+        :param count: label of the component
+        :return:
+        """
         for x in range(self.ROW):
             for y in range(self.COL):
                 if self.connected_components[x][y] is None and visited[x][y] is True:
@@ -38,7 +73,12 @@ class Graph(object):
                         KEYS.COLOR.name: label
                     }
 
-    def detect_connected_components(self):
+    def detect_connected_components(self) -> None:
+        """
+        This method detect all connected components.
+
+        :return: None
+        """
         visited = [[False for _ in range(self.COL)] for _ in range(self.ROW)]
         count = 0
         for numb in range(1, self.bins + 1):
@@ -49,10 +89,18 @@ class Graph(object):
                         self.DFS(i, j, visited, g)
                         self.label_connected_components(visited, numb, count)
                         count += 1
-        print(np.array(visited))
         return count
 
-    def binary_matrix(self, number, g, r, c):
+    def binary_matrix(self, number, g, r, c) -> np.array:
+        """
+        Turn the matrix into a given binary matrix to detect connected components.
+
+        :param number: how many colors labels are in the matrix
+        :param g: the graph
+        :param r: row
+        :param c: column
+        :return: matrix
+        """
         adjacency_matrix = copy.deepcopy(g)
         for i in range(r):
             for j in range(c):

A2/README.md

+# Ex 2
+
+Es wäre möglich die die coherenten und incoherenten Pixel über den Farbraum RGB zu berechnen.
+
+Daraus könnten dann Histogramme erzeugt werden.
+Jeder Bin repräsentiert dann ein Farbe und die Information "coherent" bzw. "incoherent".
+
+Die modifizierten CCV sehen dann bspw. so aus:
+
+	RotCoherent: 10,
+	RotIncoherent: 1,
+	GrünCoherent: 1,
+	GrünIncoherent: 10,
+	BlueCoherent: 5,
+	BlueIncoherent: 5
+
+Anschließend können solche Histogramme mit der "Earth Mover Distance" verglichen werden.
\ No newline at end of file