Discipline: Computer Science
This study is all about an image-based template matching recognition using the region of interest method and phase correlation algorithm. Template matching recognition is a technique in digital image processing of finding small parts of an image which match a template. In order to determine whether both images are exactly match or not, the process will be done with the use of the two methods. The first one is the use of region of interest method to get only a portion of an image that will be processed in template matching. Region of interest is very much useful for cropping an object from an image to perform template matching within sub image. The second one is the use of the phase correlation method to identify the similarity between the student output and the reference image. In using phase correlation both images should be equal in size.
The researchers concluded that combined methods of region of interest and phase correlation that is incorporated in the application is effective and produces good result. The results in comparing the traditional way of checking and the application are not a hundred percent accurate. Some of the results coming from the teacher are approximated by the application but some are not.