Could you explain appearance based representation in compute

Could you explain appearance based representation in computer vision

Solution

Appearance based approaches may be classified according to the degree of structure that is extracted from the image data.At one extreme, some techniques store explicit collection of pictures as object models and use this representation to recognize the object by either correlating an input image directly with the model images or by comparing the input image with the principal components of the model data set.

1. Edge matching

2. Divide-and-Conquer search

3. Greyscale matching

4. Gradient matching

5. Histograms of receptive field responses

6. Large modelbases

Although the various appearence based approaches to computer vision may differ substantially, they are all based on the premise that the object centered three dimensional object descriptions are not required.