app.py

# -*- coding: utf-8 -*-
"""Image Conversion App .ipynb

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1GA79rJr1_J5bHa7sWxF1GtDqqaTznVia

# ***Image Conversion from Simple image***
> To 
1. Edge Image
2. Pencile sketch
3. Painting
4. Cartoon image

###Step 1:- Import All libraries 
Namely
- Numpy (For storing Image)
- Matplotlib (For Display image)
- OpenCV (for Converting the image)
"""

# installing all imp libs to the machine
#!pip install opencv_contrib_python

import cv2
import numpy as np
import matplotlib.pyplot as plt
import gradio as gr

"""###Step 2:- Import and show the image
for this we will be using opencv and matplotlib

"""

def read_image(Image):
  '''
  This Function is made to take image input from the user
  Input: Path of the Image
  Output: Image in RGB format
  '''
  img = cv2.imread(Image)
  img = cv2.cvtColor(img,cv2.COLOR_BGR2RGB) # As opencv read the image as BGR so we have to convert it to RGB
  plt.imshow(img)
  plt.show()
  return img

filename ="/content/to.jpeg"
Image= read_image(filename)

"""### Step 3:- Converting the Image to respective types
Making diffrent Function for each work

####4. Cartoon Image
"""

# create Edge
def edge_mask(Image, line_size, Blur_value):
  grey_img=cv2.cvtColor(Image,cv2.COLOR_RGB2GRAY)
  grey_blur= cv2.medianBlur(grey_img, Blur_value)
  edge_image= cv2.adaptiveThreshold(grey_blur, 255, cv2.ADAPTIVE_THRESH_MEAN_C, cv2.THRESH_BINARY, line_size, Blur_value)
  return edge_image
# line_size,Blur_value=7,7
# edges= edge_mask(Image,line_size,Blur_value)
# plt.imshow(edges, cmap="binary")
# plt.show()

#reduce Colour Palet
def colour_quantization(Image, k):
  data=np.float32(Image).reshape(-1,3)
  critria= (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 20, 0.01)
  ret, label, center= cv2.kmeans(data,k, None, critria, 10, cv2.KMEANS_RANDOM_CENTERS)
  centers= np.uint8(center)
  result=centers[label.flatten()]
  result=result.reshape(Image.shape)
  return result
# Paint= colour_quantization(Image,10)
# plt.imshow(Paint)
# plt.show()
# cv2.imwrite("final1.jpg",Paint)

def cartoon(Paint,edges):
  Blurred_img= cv2.bilateralFilter(Paint, d=8 ,sigmaColor=200, sigmaSpace=200)
  Final=cv2.bitwise_and(Blurred_img,Blurred_img,mask=edges)
  # plt.imshow(Final)
  # plt.show()
  # Final= cv2.cvtColor(Final, cv2.COLOR_RGB2BGR)
  # cv2.imwrite("Final.jpg", Final)
  return Final
# cartoon()

def Main_cartoon(Image,Line_size,Blur_value,Color_count):
  edge_mask_img= edge_mask(Image, Line_size, Blur_value)
  Paint_img=colour_quantization(Image, Color_count)
  cartoon_img= cartoon(Paint_img,edge_mask_img)
  # cartoon_img=cv2.cvtColor(cartoon_img, cv2.COLOR_RGB2BGR)
  return cartoon_img

x=Main_cartoon(Image,7,7,10)

"""####3. Painting"""

def Painting(Image, Colour_count):
  Painting=colour_quantization(Image,Colour_count)
  # Painting=cv2.cvtColor(Painting, cv2.COLOR_RGB2BGR)
  return Painting

y=Painting(Image,10)

"""####2.Pencil Sketch"""

def Pencil_sketch(Image):
  grey_img=cv2.cvtColor(Image, cv2.COLOR_BGR2GRAY)
  invert_img=cv2.bitwise_not(grey_img)
  blur_img=cv2.GaussianBlur(invert_img, (111,111),0)
  invblur_img=cv2.bitwise_not(blur_img)
  sketch_img=cv2.divide(grey_img,invblur_img, scale=256.0)
  return sketch_img

z=Pencil_sketch(Image)

"""####1.Edge Sketch"""

def Edge_sketch(Image):
  edge_img = cv2.Canny(Image,100,200)
  edge_img = cv2.cvtColor(edge_img,cv2.COLOR_BGR2RGB)
  return edge_img

t=Edge_sketch(Image)

"""###Step 4:- Testing"""

plt.imshow(x)
plt.show()
cv2.imwrite("Final.jpg", x)

plt.imshow(y)
plt.show()
cv2.imwrite("final1.jpg",y)

plt.imshow(z)
plt.show()
cv2.imwrite("final2.jpg",z)

plt.imshow(t)
plt.show()
cv2.imwrite("final3.jpg",t)

"""### Step 5:- Gradio app and driver code

#### Driver Code
"""

def Sketch_app(Image,Type,Color_count,Line_size,Blur_value):
  if Type == "Cartoon":
    Result = Main_cartoon(Image,Line_size,Blur_value,Color_count)
  elif Type == "Painting":
    Result = Painting(Image,Color_count)
  elif Type == "Pencil sketch":
    Result = Pencil_sketch(Image)
  else:
    Result = None
  return Result

"""#### Gradio app

##### Installing and Importing Gradio
As Gradio is a 3rd party library we have to install it in Our Run time before Executing
"""

#!pip install gradio



Image_conversion = gr.Interface(
    fn = Sketch_app,
    inputs=[
            gr.Image( tool="select",label="Image to Convert", show_label=True),
            gr.Dropdown( choices = ["Pencil sketch","Painting","Cartoon"],label="Type to convert", show_label=True),
            gr.Slider( minimum=5, maximum=20, value =10, step= 1,label="Number of colour to be used in photo ( use only in case of painting and cartoon)", show_label=True),
            gr.Slider( minimum=5, maximum=10, value =7, step= 1,label="Blurr effect to be used in photo ( use only in case of cartoon)", show_label=True),
            gr.Slider( minimum=5, maximum=10, value =7, step= 1,label="Thickness of edges to be used in photo ( use only in case of cartoon)", show_label=True)],
    outputs= "image",
    tittle = " Image Conversion App",
    description = """This is an image conversion app we take a regular photo and convert it into Cartoon, Painting, Pencil sketch. it is purely python based used 
    Gradio(for interfaceing), OpenCV (For image conversion), Numpy(for storing the image), Matplotlib (for displaying the image) """,
    theme = "dark"
)

Image_conversion.launch()