CS 143 / Project 1 / Image Filtering and Hybrid Images
Example of an Hybrid image.
Hybrid Image
A hybrid image is an image generated by superimposing two images with different spatial scales. One image is passed through a low pass filter and another image is passed through a high pass filter. Thus taking all the low frequencies from one image and high frequencies from another image we can generate a hybrid image.
A hybrid image is obtained by combining two images IM1 and IM2 , one filtered with a low-pass filter G1 (Gaussian) and the second one filtered with a high-pass filter (1 - G1) : H = IM1 * G1 + IM2 * (1 - G2 ). Hybrid image is defined by two frequencies, one is cut-off frequency for low pass image and other is cut-off frequency for high pass image. In this project I have used same cut-off frequency for two different Gaussian.
Perception of Hybrid image
The Motivation behind hybrid images is the multi-scale processing of human visual system. The images appear to change based on the viewing distance. For example look at the above image. It looks like a cat when you look closely but when you move away from the screen it looks like a cute dog.
Results
|
|
|
|
Other Results
|
|
In above results we can see that when viewing images closely it looks like one image but when we move away or look at images from a distance the we see a different images. Hence perception of image is the function of the distance as well as the screen size.