Semantic Content Recognition and Scene Matching

Abstract: Scene matching is the process of locating a region of an image with the corresponding region of another image

where both the image regions represent the same scene. Classic scene matching methods can be divided into two major groups:

area- based algorithms that used image pixel intensity values and feature - based algorithms that utilized low-level and high-level

features.

Simulation results:

Fig.1. Scene Matching with SIFT features.

.

Last updated: June 8, 2009