Contents
INTERPLAY OF IMAGES AND FILTERS
Past summer i was playing around with images using adobe photoshop,making it sharp in contrast,adding a bit noise here and there,making it look blurry,making the image pretty by the photo retouching tools in photoshop.But it never occurred to me ,the basis behind those applications,what made the images look the way i wanted them to? How did images look so bright? Until Professor Boutin introduced the topic of "Filters" in her ece 438 class,i had never thought of images the way i do now,before.
Two basic filters used are 1> Average Filter
2> Edge Detector
Average Filter
This filter basically takes the average of 2 neighboring pixels and computes its mean which automatically results in the elimination of unneccessary noise.The final image looks 'smooth' due to the average filter.The above filter works on the following principle:
$ \ \left ( \frac{\mathbf{y }\left [ \mathbf{n } \right ]=left ( \frac{\mathbf{x }\left [ \mathbf{n } \right ]+ \mathbf{x }\left [ \mathbf{n} - \mathbf{1} \right ]} {2} \right ) $
Edge Detector
Egde is a jump in intensity from one pixel to the next.Edge is an important tool in image processing,because it gives information about the surface,depth properties.In an image when one traverses from a region of similar pixels of low frequency to a region of high frequency,an edge is encountered at intersection of the two regions.The equation which formulates an edge detector is:
$ \ \left ( \frac {\mathbf{x }\left [ \mathbf{n } \right ]- \mathbf{x }\left [ \mathbf{n} - \mathbf{1} \right ]}{2}\right ) $
In practice many approaches are used to detect edges in images.One method constitutes computing the maxima and minima of the first derivative of the image.The other method is called 'Laplacian method',which finds the zero order crossings of the second derivative of the image.
An application of the above filters is as follows :
Inference
A smooth image is obtained by applying the average filter as seen in the above figure.In case of the edge detected image, a comparison of neighboring pixels occurs.If the comparison leads to zero,then that contributes to the black parts of the output image.When the comparison leads to a finite number,then that depicts the white parts of the image,which represents the edge.
References
1. http://www.thewashingtonnote.com/archives/2009/06/twn_on_road_lon/
2. http://en.wikipedia.org/wiki/Edge_detection
--Apanja 21:55, 13 October 2009 (UTC)Ananya Panja
