Merge pull request #165 from Namyalg/Image_filters

Image filters using 3-D convolution

Author: powerexploit

Image-Processing/Image_filters_using_convolution/3D_convolve.py

@@ -0,0 +1,73 @@
+#This script implements convolution over images.
+#This can be used to apply filters to images.
+
+#Imports
+
+import numpy as np
+from PIL import Image
+import cv2
+
+#size of image
+width = 300
+height = 360
+
+#kernels or image filters
+#This is to blur the image, experiment with different values, each of the streams can have different values
+'''b_kernel = np.array([0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11]).reshape(3,3)
+r_kernel = np.array([0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11]).reshape(3,3)
+g_kernel = np.array([0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11 , 0.11]).reshape(3,3)
+'''
+print("Enter 9 values for each of the streams")
+b_kernel = np.array(list(map(int, input("Enter the kernel to be applied to the blue stream").split()))).reshape(3,3)
+r_kernel = np.array(list(map(int, input("Enter the kernel to be applied to the red stream").split()))).reshape(3,3)
+g_kernel = np.array(list(map(int, input("Enter the kernel to be applied to the green stream").split()))).reshape(3,3)
+
+#Image is read and resized
+img = cv2.imread('path to image')
+img = cv2.resize(img, (width, height))
+
+#The pixels are extracted from the image
+pixels = []
+for i in range(height):
+     for j in range(width):
+          pixels.append(img[i, j])
+           #The RGB streams are stored in different arrays to facilitate the process of convolution
+           b = []
+           g = []
+           r = []
+           for i in pixels:
+              b.append(i[0])
+              g.append(i[1])
+              r.append(i[2])
+
+#Each of the streams are resized
+r = np.array(r).reshape(height, width)
+b = np.array(b).reshape(height, width)
+g = np.array(g).reshape(height, width)
+
+#function to apply convolution or multiply kernels
+def convolute(arr, height, width, kernel):
+      convoluted_matrix = []
+      for i in range(1, height-2):
+          for j in range(1, width-2):
+            temp = arr[i:i+3, j:j+3]
+            prod = np.multiply(temp, kernel)
+            convoluted_matrix.append(np.sum(prod))
+
+      convoluted_matrix = (np.array(convoluted_matrix).reshape(height-3, width-3))
+      return(convoluted_matrix)
+
+r_convol = convolute(r, height, width, r_kernel)
+g_convol = convolute(g, height, width, g_kernel)
+b_convol = convolute(b, height, width, b_kernel)
+
+width = width - 3
+height = height - 3
+combine = np.zeros([height, width, 3], dtype=np.uint8)
+for i in range(height):
+      for j in range(width):
+            combine[i, j] = [int(r_convol[i][j]), int(g_convol[i][j]), int(b_convol[i][j])]
+
+#The 3 matrices or streams are combined, and stored
+img = Image.fromarray(combine)
+img.save('Funkyfilter.jpg')

Image-Processing/Image_filters_using_convolution/README.md

@@ -0,0 +1,30 @@
+
+## Image filters using convolution ##
+
+- Convolution is a technique widely used in image processing . Convolution is a mathematical operation on two functions that produces a third function expressing how the shape of one is modified by the other. (As defined in Wikipedia).
+
+- Though this technique is widely used in image processing, in my implementation it can be used as an image filter.
+
+- Images are made of pixels, colored images have 3 streams (red, green, blue), these can be visualised as a matrix which is 3 dimensional
+- Each of the red, blue and green are streams, stacked together to form a cube
+
+- A kernel or a smaller matrix, in this case a 3X3 matrix is used to perform convolution.
+
+- The kernel is used to perform dot product with the image. The resultant is an image which has different properties.
+
+- This script written in Python, implements convolution over colored images.
+
+- The end result is the application of filters to images.
+
+- I have used tensorflow, numpy and the image processing library.
+
+- The images attached show an image before and after blurring.
+
+![Image](./images/before.png)
+
+- This is before applying convolution.
+
+![Image](./images/after.png)
+
+- This is after applying convolution.
+