engleski

Template Matching Technique for Image Object Recognition

The automatic object recognition is important but challenging task in computer vision. The main difficulty is that the system must be able to differentiate specific object in spite of restrictive factors such as lighting conditions or different object orientations and positions with respect to the camera. The most commonly used method when trying to recognize a specific object, given in a form of a template, as a part of a target image is template matching technique. In this work, object recognition is implemented as a two-stage process. First, histogram matching is applied in order to find areas where object is most likely to be located. However, it is shown that using only histogram as a similarity criterion is not sufficient. Therefore, best-matching subparts of a scene image are extracted and further analyzed in frequency domain. Our experiment shows that object is successfully recognized on various different scenes. Furthermore, even if object has already been partially recognized during the first-step histogram-based template matching, combination with the second-step analysis in Fourier domain can additionally enhance the recognition process.

object recognition ; template matching ; histogram ; spectrum

nije evidentirano