Image Processing: Understanding Image Representation, Acquisition, Display, and Errors, Slides of Computer Graphics

Download Image Processing: Understanding Image Representation, Acquisition, Display, and Errors and more Slides Computer Graphics in PDF only on Docsity!

Image Processing

Overview ^ Image Representation^ ^ What is an image? ^ Halftoning and Dithering^ ^ Trade spatial resolution for intensity resolution^ ^ Reduce visual artifacts due to quantization ^ Sampling and Reconstruction^ ^ Key steps in image processing^ ^ Avoid visual artifacts due to aliasing

What is an Image? ^ An image is a 2D rectilinear array of pixels

Continuous image

Digital image

What is an Image? ^ An image is a 2D rectilinear array of pixels

Digital image A pixel is a sample, not a little square!!

Continuous image

Image Display ^ Re-create continuous function from samples^ ^ Example: cathode ray tube

Image is reconstructedby displaying pixels withfinite area (Gaussian)

Image Resolution ^ Intensity resolution^ ^ Each pixel has only “Depth” bits forcolors/intensities ^ Spatial resolution^ ^ Image has only “Width” x “Height” pixels ^ Temporal resolution^ ^ Monitor refreshes images at only “Rate” Hz

Quantization ^ Artifact due to limited intensity resolution^ ^ Frame buffers have limited number of bits perpixel^ ^ Physical devices have limited dynamic range

Blue channelGreen channel Red channel

Uniform Quantization^ ^

^

^ 

trunc,

^

yxI

yxP

I(x, y) P(x, y)2 bits per pixel

Reducing Effects of Quantization ^ Halftoning^ ^ Classical halftoning

Classical Halftoning ^ Use dots of varying size to representationintensities^ ^ Area of dots proportional to intensity in image

I(x, y)^

P(x, y)

Halftone Patterns ^ Use cluster of pixels to represent intensity^ ^ Trade spatial resolution for intensity resolution

Halftone Patterns ^ How many intensities in a

n^ x^ n

cluster?

Sampling and Reconstruction

Aliasing ^ In general:^ ^ Artifacts due to under-sampling or poorreconstruction ^ Specifically, in graphics:^ ^ Spatial aliasing^ ^ Temporal aliasing

Under-sampling