Segmentation Using Canny+Watershed in Opencv-python

Python Script is given below:

import cv2

# Importing opencv library
import numpy as np
#  Importing NumPy,which is the fundamental package for scientific computing with Python
img = cv2.imread('C:\\Users\Ram\Pictures\\humans\\368078.jpg')
# Read the image from disk
cv2.imshow("Original image", img)  # Display image
img_float = np.float32(img)  # Convert image from unsigned 8 bit to 32 bit float
criteria = (cv2.TERM_CRITERIA_EPS+cv2.TERM_CRITERIA_MAX_ITER, 10, 1)
# Defining the criteria ( type, max_iter, epsilon )
# cv2.TERM_CRITERIA_EPS - stop the algorithm iteration if specified accuracy, epsilon, is reached.
# cv2.TERM_CRITERIA_MAX_ITER - stop the algorithm after the specified number of iterations, max_iter.
# cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER - stop the iteration when any of the above condition is met.
# max_iter - An integer specifying maximum number of iterations.In this case it is 10
# epsilon - Required accuracy.In this case it is 1
k = 50  # Number of clusters
ret, label, centers = cv2.kmeans(img_float, k, None, criteria, 50, cv2.KMEANS_RANDOM_CENTERS)
# apply kmeans algorithm with random centers approach
center = np.uint8(centers)
# Convert the image from float to unsigned integer
res = center[label.flatten()]
# This will flatten the label
res2 = res.reshape(img.shape)
# Reshape the image
cv2.imshow("K Means", res2)  # Display image
cv2.imwrite("1.jpg", res2)  # Write image onto disk
meanshift = cv2.pyrMeanShiftFiltering(img, sp=8, sr=16, maxLevel=1, termcrit=(cv2.TERM_CRITERIA_EPS+cv2.TERM_CRITERIA_MAX_ITER, 5, 1))
# Apply meanshift algorithm on to image
cv2.imshow("Output of meanshift", meanshift)
# Display image
cv2.imwrite("2.jpg", meanshift)
# Write image onto disk
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# Convert image from RGB to GRAY
ret, thresh = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY+cv2.THRESH_OTSU)
# apply thresholding to convert the image to binary
fg = cv2.erode(thresh, None, iterations=1)
# erode the image
bgt = cv2.dilate(thresh, None, iterations=1)
# Dilate the image
ret, bg = cv2.threshold(bgt, 1, 128, 1)
# Apply thresholding
marker = cv2.add(fg, bg)
# Add foreground and background
canny = cv2.Canny(marker, 110, 150)
# Apply canny edge detector
new, contours, hierarchy = cv2.findContours(canny, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
# Finding the contors in the image using chain approximation
marker32 = np.int32(marker)
# converting the marker to float 32 bit
cv2.watershed(img,marker32)
# Apply watershed algorithm
m = cv2.convertScaleAbs(marker32)
ret, thresh = cv2.threshold(m, 0, 255, cv2.THRESH_BINARY+cv2.THRESH_OTSU)
# Apply thresholding on the image to convert to binary image
thresh_inv = cv2.bitwise_not(thresh)
# Invert the thresh
res = cv2.bitwise_and(img, img, mask=thresh)
# Bitwise and with the image mask thresh
res3 = cv2.bitwise_and(img, img, mask=thresh_inv)
# Bitwise and the image with mask as threshold invert
res4 = cv2.addWeighted(res, 1, res3, 1, 0)
# Take the weighted average
final = cv2.drawContours(res4, contours, -1, (0, 255, 0), 1)
# Draw the contours on the image with green color and pixel width is 1
cv2.imshow("Watershed", final)  # Display the image
cv2.imwrite("3.jpg", final)  # Write the image
cv2.waitKey()  # Wait for key stroke

The result of the above script is given in the pdf below (2nd Answer):