0

I am currently working on a project using Python to detect Snapcodes among a set a pictures. For those who don't know what a Snapcode is, here are some examples.

I am looking for the most simple solution to perform this detection using the Python language.

Dev-iL
  • 23,742
  • 7
  • 57
  • 99
Matt
  • 1
  • Are they pictures of the codes taken with a camera? Or are they images with codes in them? – Blender Apr 16 '16 at 17:43
  • No they are screenshot of a Snapcode exactly like that http://cdn0.dailydot.com/uploaded/images/original/2015/7/1/snapchat-qr-selfie.jpg and they are among a set of photographies pictures. Thanks Blender. – Matt Apr 16 '16 at 17:48
  • Then all you need to do is find roughly-square yellow blobs in the image and filter them out based on the presence of dots. – Blender Apr 16 '16 at 17:49
  • Well, the location may differ because of the size of the screen. Moreover sometimes screenshot are cropped. So your solution may not work. Thanks :) – Matt Apr 16 '16 at 17:51
  • As in the actual code is partially obstructed? – Blender Apr 16 '16 at 17:52
  • Yes, the snapcodes screenshots may vary, but the snapcode is not obstructed. – Matt Apr 16 '16 at 17:54
  • 1
    So just do yellow blob detection, find the ones with the right shape, and you're good to go. – Blender Apr 16 '16 at 18:28
  • Hi Matt. How did this go? I'm trying to do something similar. – Chaz Oct 11 '16 at 16:46

1 Answers1

1

You can use an image classification algorithm like svm to find the Snap codes from a given set of images.In the case of svm you first need to train svm using some pictures of snap codes and some other non-snap code pics.Once training phase completed you can predict whether given image is a snapcode or not.SVM is already implemented in Opencv.

Usefull stackoverflow question

what is svm with code-but in java

opencv implementation of image classifier using svm with python code

Community
  • 1
  • 1
Emmanu
  • 749
  • 3
  • 10
  • 26