Image Morphing

Image Morphing

Presented here is an investigation into the technique of morphing: a method of creating smooth transitions between two images. Morphing is a common technique in special effects for entertainment, but is becoming important in the fields of computer vision and psychology. Mesh warping and crossfading techniques are employed to morph between various still images. Morphing between full-motion film sequences is also considered and a basic example included. The algorithms used are described in detail including an analysis of the number of ‘critical points’ that must be identified for a successful morph and the possible problems that could be encountered. The mesh warping method, despite being one of the oldest techniques, is shown to be very effective.
Image Morphing Report
Matlab Code
Image 1
51 frame morph - 5 seconds duration - 60 feature locator points
Avi Video - xvid
Image 1
151 frame morph - 6 seconds duration - 114 feature locator points
Avi Video - xvid
Image 1
31 frame video morph - 3 seconds duration - ~40 consistently placed locator points on all frames
Avi Video - xvid
Image 1
101 frame morph - 5 seconds duration -60 feature locator points
Avi Video - xvid
Image 1
Template Interface Video - 2:09 minutes - 3.02 MB Outputs ImageFileName.met (MatLab data file containing the original image and a vector of points)
Avi Video - xvid
Image 1
Dual Interface Video - 1:02 minutes - 1.89 MB - 10 Point selection on two facing images Outputs 2x ImageFileName.met (MatLab data file containing the original image and a vector of points)
Avi Video - xvid
Image 1
Morphing Video - 1:55 minutes - 2.12 MB - 30 Frames of Morph* Reads ImageFileName.met (MatLab data file containing the original image and a vector of points)
Avi Video - xvid