Image Histograms and applications - Computer Science at Rutgers

98 downloads 85 Views 3MB Size Report
Outlines. ▫ Digitizing images. ▫ Image histograms and its applications. ▫ Sources: ▫ Burger and Burge “Digital Image Processing” Ch. 4. ▫ Digitizing images.
CS443: Digital Imaging and Multimedia Histograms of Digital Images

Spring 2008 Ahmed Elgammal Dept. of Computer Science Rutgers University

Outlines ƒ Digitizing images ƒ Image histograms and its applications

ƒ Sources: ƒ Burger and Burge “Digital Image Processing” Ch. 4

1

Digitizing images ƒ What is projected on the image plan is a distribution of light energy that is: ƒ ƒ ƒ

Two-dimensional Time-dependent Continuous

ƒ To go digital: ƒ Spatial sampling ƒ Temporal sampling ƒ Quantization of pixel values

ƒ Digital image: two-dimensional, ordered matrix of integers, i.e., a two-dimensional function of integer coordinates NxN that maps a range of image values P

2

ƒ Image resolution: number of image elements per measurement. ƒ Image coordinate system

3

Image Histograms ƒ Histograms are used to depict image statistics in an easily interpreted visual format ƒ Useful during image capturing: now already in digital cameras ƒ Used to improve the visual appearance of an image ƒ Can also be used to determine what type of processing has been applied to an image.

ƒ Image histogram: describes the frequency of the intensity values that occur in an image

4

ƒ Histograms don’t encode information about the spatial arrangement of pixels in the image ƒ We cannot reconstruct an image given only it’s histogram

5

Interpreting Histograms ƒ Histograms depicts problems that originate during image acquisition ƒ Exposure, contrast, dynamic range

ƒ Histograms can be used to detect a wide range of image defects: saturation, spikes and gaps, impact of image compression

Histograms and Exposure

6

Histogram and Contrast

Dynamic Range

Dynamic Range: the number of distinct pixel value in an image

7

Detecting Image Defects ƒ There is no ideal or optimal histogram shape. It depends on the image and on the application ƒ Image Defects: ƒ Saturation: the illumination values lying outside of the sensor’s range are mapped to its maximum or minimum values: spike at the tails ƒ Spikes and Gaps in manipulated images. Why? ƒ Impact of image compression

ƒ Histograms show the impacts of image compression ƒ Ex: in GIF compression, the dynamic range is reduced to only few intensities (quantization)

8

ƒ Ex: JPEG compression on a line graphics.

JPEG compression

9

Computing Histograms

ƒ Histograms of images with more than 8 bits: ƒ Binning

ƒ Ex: B=256 for 14 bit image K=16384, bin width = 64

10

Color Image Histograms

Color Image Histograms ƒ For color images, two kind of histograms: ƒ Intensity histogram ƒ Individual Color Channel Histograms

ƒ Both provides useful information about lighting, contrast, dynamic range and saturation effects for individual color components ƒ They provide no information about the actual color distribution!

11

Cumulative Histograms

12

Suggest Documents