|
First attempt at using gaussian vector averaging on the zebra image. | |||
| This is the thresholded large eigenvale (region which has stripes). Even though parameter choice was random we still get a connected object that is completely separate from background. |
|
|||
| This is the edge of the segmented region and the angle image. They could also be useful for certain tasks. |  |
|
||
|
Conclusion: We found that the technique of vector averaging
works for certain tasks and found the effect of noise and size of
gaussian filter on performance. This technique works best for images which have high resolution compared to level of detail.
|
||||