import cv2 #
import numpy as np

image = '459x600-monalist.PNG'
img = cv2.imread(image)
cv2.imshow('Mona Lisa',img)
cv2.waitKey( 0 )

smooth1_img = cv2.GaussianBlur( img, (7,7), 0 )
cv2.imshow('Mona Lisa', smooth1_img)
cv2.waitKey( 0 )

# more blurring
smooth2_img = cv2.GaussianBlur( img, (23,23), 0 )
cv2.imshow('Mona Lisa', smooth2_img)
cv2.waitKey( 0 )

# in class - 
# convert image to a different color space - here we
# convert from RGB to grayscale
gray = cv2.cvtColor( smooth2_img, cv2.COLOR_BGR2GRAY )
cv2.imshow('Mona Lisa', gray )
cv2.waitKey( 0 )

# we will do edge detection later in the semester here is
# a module, Canny, in openCV that does just that on a
#	grayscale image 'img'
# it has two parameters in addition to the image img. 'lower'
# and 'upper'.
#	 Canny( img, lower, upper )
# lower and upper between [0,255] are just thresholds on how sensitive you are
# to detecting edges, for now we will use, 30 and 100.
edges = cv2.Canny( gray, 30, 100 )
cv2.imshow('Mona Lisa', edges )
cv2.waitKey( 0 )


# for cropping, we did it directly in class, here in this
# example we will define the arguments of cropping via
# x, y coordinates to make things more clear, typically,
# the origin starts at the top left of an image, so that is 
# (0,0). The y coordinate system then goes 'downwards' so that
# is the only really confusing part of displaying images in
# openCV.
# shape[1] x shape[0] is the dimension of the original image
print "Dimension of the image are ( %d, %d )" % (img.shape[1], img.shape[0])
# h = denotes height
# w = denotes width
y1 	= img.shape[0]/4 - 130
h 	= 260
y2	= y1+h

x1 	= img.shape[1]/2 - 100
w 	= 220
x2	= x1+w;

# here is a cropped image this will at least give you a start
# to complete the exercises some more hints further below
# on that.

# slightly different from class in that we will assign the variables
# above to make the slicing more apperant
img_cropped = img[  y1:y2, x1:x2 ]
cv2.imshow('cropped', img_cropped )
cv2.waitKey(0)

# HINTS: to select a region of interest (roi) in an image you can use the idea
# of cropping, and store it to a new image variable
# then after defining and tranforming this roi, you can then copy it
# to the original image.
#
# 1) roi = ... define it similar to the slicing idea above ...
# 2) tranform the roi
# 3) set a new image (e.g., newImage_to the old untranformed image)
# 4) newImage[ ... ] = transformedimage
#
cv2.destroyAllWindows()