“Hybrid” Image (Framed Image) Detection

I’m trying to implement an “hybrid” image tracker. This technique has the objective to detect certain type of images to enable augmented reality. Here are early results!

What’s an Hybrid Image?

An “hybrid image” – also called “framed image” – is an image that is framed with a predefined geometric shapes. The term “hybrid” is here to indicate that the tracking acts between marker tracking and natural image tracking principles. Most of the time, the “frame” is constructed using a black rectangle, but it could be rectangles, circles, in fact every shape that is easy to track.

Here are some sample i used:

Hybrid Image constructed with a movie database

Hybrid Image constructed with a movie database

Nike Commercial - Hybrid Image

Nike Commercial enhanced with round shapes to ease detection

Video sample:

I worked on an algorithm that uses the black frame to ease the recognition of up-to-500 hybrid images. Here is the result:


2 thoughts on ““Hybrid” Image (Framed Image) Detection

  1. Hi,

    You can use a totally standard approach to detect the black borders see for example opencv’s border detection algorithm.

    Recognition is also very simple, you can use SIFT (but CPU consuming). ORB/BRIEF/FERNS are better descriptor (near as robust and very fast). Please follow our work (xzimg.com) since a free version of the SDK is soon available.


Leave a Reply

Your email address will not be published. Required fields are marked *


× two = 12

* Copy This Password *

* Type Or Paste Password Here *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>