Image recognition using Python

create_example_file.py

def createExamples():
	createFile = open('exFile.txt','a')
	numberEx = range(0,10) #0.1,1.1,2.1
	numberVer = range(1,10) #0.1,0.2,0.3
	
	for eachNum in numberEx:
		for eachVer in numberVer:
			imgFilePath = 'images/numbers/'+str(eachNum)+'.'+str(eachVer)+'.png' #create image path
			i = Image.open(imgFilePath)
			iar = np.array(i)
			iar = str(iar.tolist())

			lineToWrite = str(eachNum) +'::'+iar+'\n' #generating liine to write in file
			createFile.write(lineToWrite) # writing to file

      
#createExamples() //function call

find_the_number.py

def whatNumberIsThis(imgFilePath):
	matchedAr = []
	loadExmp = open('exFile.txt','r').read()
	loadExmp = loadExmp.split('\n')

	exImage = Image.open(imgFilePath)
	eIAr = np.array(exImage)
	eIAr = threshold(eIAr)
	eiArL = eIAr.tolist()
	
	imgToFind = str(eiArL)
	
	for eachExmp in loadExmp:
		if len(eachExmp) > 3 :
			eachExmp = eachExmp.split('::')
			numExmp = eachExmp[0]
			numAr = eachExmp[1]
			
			eachPix = numAr.split('],')
			eachPix_imgToFind = imgToFind.split('],')
			
			x=0;
			while(x < len(eachPix)):
				if eachPix[x] == eachPix_imgToFind[x]:
					matchedAr.append(int(numExmp))
				x += 1		
	c=Counter(matchedAr)
	print(c)

lesson_2_notes.py

import numpy as np
from PIL import Image 
#(pil is pillow module)

i = Image.open('images/dot.png')

iar = np.asarray(i)

print(iar)

'''
install python and pip (as pip contains all the required modules like numpy,matplotlib and pillow or pil)
* start python using terminal and import numpy matplotlib and PIL
* import Images from PIL(run 'from PIL import Image')


So each pixel is measured in RBGA, so an example row is [255, 255, 255, 255], what does that mean?

This means we're looking at a 256-color image, since programming starts with a 0 rather than a 1.
This color means 255 red, 255 green, 255 blue, and then 255 Alpha.

this will print an array of RGB-A(red,green,blue and alpha )pixels on the screen.
Alpha is a measure of how opaque an image is. The higher the number, the more solid the color is, 
the lower the number, the more transparent it is.
'''

threshold_function.py

import numpy as np
import matplotlib.pyplot as plt
import time
from PIL import Image

def threshold(imageArray):
	balAr = []
	newAr = imageArray
	from statistics import mean
	for eachRow in imageArray:
		for eachPix in eachRow:
			avg = mean(eachPix[:3])
			balAr.append(avg)
	balance = mean(balAr)

	for eachRow in newAr:
		for eachPix in eachRow:
			if mean(eachPix[:3]) > balance :
				eachPix[0] = 255
				eachPix[1] = 255
				eachPix[2] = 255
				eachPix[3] = 255
			else:
				eachPix[0] = 0
				eachPix[1] = 0
				eachPix[2] = 0
				eachPix[3] = 255
	return newAr		

i1 = Image.open('images/dot.png')
i1_ar = np.array(i1)
i1_ar = threshold(i1_ar)

i2 = Image.open('images/numbers/0.1.png')
i2_ar = np.array(i2)
i2_ar = threshold(i2_ar)

i3 = Image.open('images/numbers/y0.5.png')
i3_ar = np.array(i3)
i3_ar = threshold(i3_ar)

i4 = Image.open('images/dotndot.png')
i4_ar = np.array(i4)
i4_ar = threshold(i4_ar)

fig = plt.figure()

ax1 = plt.subplot2grid((8,6),(0,0),rowspan = 4,colspan = 3)
ax2 = plt.subplot2grid((8,6),(4,0),rowspan = 4,colspan = 3)
ax3 = plt.subplot2grid((8,6),(0,3),rowspan = 4,colspan = 3)
ax4 = plt.subplot2grid((8,6),(4,3),rowspan = 4,colspan = 3)

ax1.imshow(i1_ar)
ax2.imshow(i2_ar)
ax3.imshow(i3_ar)
ax4.imshow(i4_ar)

plt.show()