The first demo seeks to illustrate the way in which the 3D information from a stereo pair of images is differenced, pixel by pixel, before being strongly blurred and mixed with the basic original low resolution information stored in the B/Y vision channel for onward transmission to the central cortex. For each stereo image pair the composite 3D is first shown in the form of a grey anaglyph, where red and green patches indicate the local parallax existing between the left & right images. In these grey anaglyphs the red & green bands are associated with edges in a way which codes both the magnitude of the parallax and whether the local portion of the scene is closer or further distant than the plane of image fusion.
This is then followed by two difference images, where the positive & negative partial differences are split In the demo it should be noted that the difference images are only presented in an unblurred form, since the actual local levels of the differences are very small and would not readily be discernable when blurred. Furthermore they are shown as negatives (that is, against a white background) in order to permit ready visualisation (which would be very difficult to see even for sharp representations if viewed against a black background!).
Finally the resultant NE3D composite image for the scene is shown - in full colour..