A simple trick to solve the problem is to resize the template image to multiple scales and then compare it to the source image. Changes in the size of the source image or template image will affect the performance of the algorithm. One of the challenges of the template-based approach is scale invariance. Both images are converted into binary images or in black and white and then template matching techniques like normalized cross-correlation, cross-correlation, and sum of squared difference are applied. The template image is moved one pixel at a time from left to right or from top to bottom to enable to calculate some numerical measure of similarity to the patch it overlaps. Simple template matching involves comparing the template image against the source image by sliding it. The template-based approach is easier to implement than feature-based.
0 Comments
Leave a Reply. |