text_image.py

import boto3
from pprint import pprint
import image_helpers

client = boto3.client('rekognition')

imgurl = 'http://docs.aws.amazon.com/rekognition/latest/dg/images/text.png'
imgurl = 'https://image.slidesharecdn.com/howtooverlaytextonimagesjpgs-141214122828-conversion-gate01/95/how-to-overlay-text-on-images-5-simple-methods-1-638.jpg?cb=1433017251'
#imgurl = 'https://lh3.ggpht.com/AIOlcNtjP5seRwQp62nUEpleulvRgvgMjTm9bR72uMQNJsJfp1O0Op1_ZXfnEYD5uA=h900'
imgurl = 'https://blog.njsnet.co/content/images/2017/02/trumprecognition.png'
imgurl = 'http://ww2.eclipseadvantage.com/images/Photos/people_resized.jpg'
# grab the image from online
imgbytes = image_helpers.get_image_from_url(imgurl)

rekresp = client.detect_text(Image={'Bytes': imgbytes})
#pprint(rekresp['TextDetections'])
imgString=""
for img in rekresp['TextDetections']:
    imgString=imgString+img['DetectedText']+" "

pprint(imgString);
half, rem = divmod(len(imgString), 2)
frontA = imgString[:half + rem]
backA = imgString[half + rem:]
pprint(frontA)
#pprint(', '.join(imgText))