1

Image Basics

Hao Jiang

Computer Science Department

Sept. 4, 2014

Image Formulation

The most common way to obtain an image is from a camera

2

A “Simple” Camera

3

Let’s hold a sensor (a film) in front of the object.Hopefully we will have an image…

A “Simple” Camera

4

Unfortunately, at the same image point, light may comefrom different source points on an object.

The Pinhole Camera

5

Camera with Lens

6

The Imaging Model

7

lighting

Surface property: material, geometry.

Camera pose,Optical properties

Images as Surfaces

Image can be treated as a 2D function z = f(x, y).

Image Profile

9

Sampling

To “digitize” the continuous image, we need to sample the image first.

Sampling on a grid Sampling problem

The image of Barbara

Aliasing due to sampling

fs = 2.5f

fs = 1.67f

0 10 20 30 40 50 60 70 80 90 100-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

t

Am

plit

ud

e

Original signal

A new component is added

This is denotedas aliasing.

0 10 20 30 40 50 60 70 80 90 100-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

t

Am

plitu

de

Image Resolution

Sensor: size of the real world scene into a single image pixel.

Image: number of Pixels.

14

Digitization

The samples are continuous and have infinite number of possible values.

The digitization process approximates these values with a fixed number of numbers.

To represent N numbers, we need log2N bits.

So, what determines the number of bits we need for an image?

Image as Matrices

16

174 167 184207 213 227

Types of Digital Images

Grayscale image Usually we use 256 levels for each pixel. Thus we need

8bits to represent a pixel (2^8 == 256) Some images use more bits per pixel, for example MRI

images could use 16bits / pixel.

A 8bit grayscaleImage.

Binary Image

A binary image has only two values (0 or 1).

Binary image is quite important in image analysis and objectdetection applications.

Gay Scale Image as a Stack of Binary Images

[ b7 b6 b5 b4 b3 b2 b1 b0]

MSB LSB

Each bit plane is a binary image.

Dithering

A technique to represent a grayscale image with a binary one.

0 1

2 3

Convert image to4 levels: I’ = floor(I/64)

Color Image

r

g

b

24 bit image

Color Table

Image with 256 colors

r

g

b

Clusters of colors

It is possible touse much less colorsto represent a color imagewithout much degradation.

Gamma Correction

Display device’s brightness is not linearly related to the input.

I’ = I

To compensate for the nonlinear distortion we need to raise it to a power again

(I’)1/ = I

for CRT is about 2.2.

Gamma Correction

Linearly increasing intensitywithout gamma correction

Linearly increasing intensitywith gamma correction

Image File Formats

An image in “ppm” format:

P6: (this is a ppm image)Resolution: 512x512Depth: 0-255 (8bits per pixel in each channel)

An image containsa header anda bunch of (integer) numbers.

Image Compression and Encoding

Raw image takes a lot of space. Compute the file sizes of a raw image that has resolution 512x512 in true color. BMP, PPM, TXT

Images can be “compressed” losslessly or lossly Lossy image format: JPEG

Losslessly compressed image format: PNG

Compression ratio and bit rate

27

Digital Video

Frame N-1

Frame 0

time

Digital video is digitizedversion of a 3D functionf(x,y,t)