Abstract
It has been well studied to automatically match images with images, or to match line drawings with line drawings. However, matching images with line drawings imposes a greater challenge. This is because the two type of data have dramatically different optical characteristics, different content, and different geometric layout. In this paper, we present a technique to automatically match images with line drawings. It uses distance transformation to covert line drawings to a feature similar to image gradient. It then uses mutual information and unconstrained nonlinear optimization to find the best location and angle for matching the images with the line drawings. Simulation on real data shows the effectiveness of our technique.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
