Visual Information Processing

builderanthologyAI and Robotics

Oct 19, 2013 (3 years and 9 months ago)

242 views



Visual Information Processing

Human Perception V.S. Machine
Perception



Human perception:


pictorial information improvement for


human interpretation




Machine perception:


scene data processing for machine


understanding.


Visual Information Processing





Why processing visual information:


-

Better perception for visualization


-

Medical imaging


-

Video images (e.g., television commercials to feature films)


-

Storage and transmission


-

Application for machine intelligence, robotics, multimedia, graphics and


human computer interaction technology




Topics in this course:


-

Image processing fundamentals


-

Image analysis


-

Motion image processing (video)


-

Object recognition


-

Application with state
-
of
-
the
-
art techniques




Basic Concepts




Computer Vision


-

simulate the human visual system, not only “see” the world, but also
“understand” the world (emulate human vision


analysis and understanding)



Image Processing


-

pre
-
process the image for better “see” or “understand”




Pattern Recognition



-

classify and recognize both image content and some other statistic data.




Computer Graphics


-

create or synthesize a virtual image



Artificial Intelligence


-

emulate human intelligence



Digital Image Processing
Development



Digital image processing:



-

Low
-
level:

Primitive operations (e.g., contrast enhancement,


sharpening in Photoshop and Photo
-
Stacker);


Image
-
>image


-

Mid
-
level:
Image segmentation, classification;


image
-
>attributes (e.g., edges, objects).


-

High
-
level:
Ensemble of recognized objects


(vision: make it understood)



1920’s
-

Digitized newspaper picture transmitted through submarine cable


(London<
-
>New York)


-

5 distinct brightness level
-
> 15 levels


Digital Image Processing
Development (Cont’d)



1960’s
-

Images from space probe (distortion correction


image
transform)



1970’s
-

Computerized Tomography (CT) (a ring of detectors collect the
x
-
rays to represent a slice)



1980’s and later


-

Computer image processing in industry, biomedical area, military
recognition, satellite imagery for weather and environment.


-

Development of signal processing.


Digital Image Processing
Development (Cont’d)

left: Original image right: Processed image


Image

Representation




Image (Monochrome image / color image)


--

f(x,y)

-

two
-
dimensional light intensity function


--

(x,y)

denote spatial coordinates;


--

the value of
f

at any point
(x, y)
is proportional to

the brightness


(or
gray level
or

gray scale
) of the image at that point.



Example:



y

x
o

Digital Image

Representation




Digital image


--

an image
f(x,y)
that has been discretized both in the image


coordinates and in brightness


--

A matrix


the elements of digital array are called


pixels (picture elements, image elements, pels)



--

In computer programming: 2D array


--

Size: width
-

number of pixels horizontally


height


number of pixels vertically



Example:



Image Display

Computer

Hardcopy

Image Processing

Software

Image Sensors

(optical to electronic)

scene

Specialized Image

Processing Hardware


(digitizer, ALU

Arithmetic logic unit)

Image Storage

Digital Image Processing System

Digital Image Processing

Fundamentals

Color Image

Processing

Image

Enhancement

Knowledge

Base

Image

Acquisition

Image

Restoration

Image

Compression

Multiresolution

Processing

Morphological

Processing

Representation

& Description

Object

Recognition

Image

Segmentation

Still Image vs.

Motion image



Still Image



JPEG, JPEG2000



All the fundamental processing



Image synthesis




Motion Image



Motion analysis and detection



Video processing and transmission (H.261, H.263, H.264, MPEG1, 2, 4, 7,
21)



Illusion

http://www.michaelbach.de/ot/fcs_thompson
-
thatcher/index.html

http://members.tripod.com/~RBHcognitions/thatc1.htm

http://www.wjh.harvard.edu/~lombrozo/home/illusions/thatcher.html



Image Analysis

-

Computer Vision




Human Vision


-

70% information from visual perception


-

30% information from sound, touch, taste, smell, ……



Computer Vision


-

object detection (edge, region, texture, color,…)


-

camera calibration


-

2D to 3D (shape from shading, shape from texture,


shape from motion, …)



Virtual Image

-

Image Synthesis



Real Image and Virtual Image:



Image analysis:
real

image


-

using computer to understand the real world


Image Synthesis:
virtual

image


-

using computer to create a virtual world (computer graphics)



Evaluation



Subjective and Objective Evaluation
:



Subjective: No better way to judge the quality of an image than human vision


-

rating



Objective: pixel
-
by
-
pixel comparison


-

mean square errors measurement


Recent Development


1.
Characteristics:





Better quality, Fast processing, Accurate Detection, and More understanding



Architecture: Parallel algorithm



Robotics and Active vision



Face and Gesture Recognition



Document Image Analysis



Texture Analysis



Motion tracking and Analysis



Color image analysis



Image Segmentation and Feature Extraction



3D Reconstruction


Relevant areas


Computer Graphics


Image Processing


Multimedia


Human Computer
Interaction/Interface



Applications


Biomedical area (CT images)


Military recognition


Satellite imagery for weather and
environment


Motion video (MPEG video)


Still image (JPEG)


TV and Film making



Stereo images as seen through LCD
Shutterglasses

Image Processing
Programming and tools



MS Windows: Visual C++



Unix, Linux, Irix: C and C++



Intel Image Processing Library.



Image Vision Library



Image format information (BMP, JPEG, TIF,…)

In this Class



You will learn:


Image Processing & Computer Vision: Basics and application


You will do:


Programming Assignments, Course Project (proposal, class
presentation, project report)



Class attendance is required.


References



Text books:


(1) Digital Image Processing (R. Gonzales)


(2) Computer Vision (L. G. Shapiro)




Journals:


IEEE Transactions on Pattern Analysis and Machine Intelligence


IEEE Transactions on Image Processing


Computer vision, Graphics and Image Processing


Pattern Recognition



Conferences:


IEEE International Conf. on Computer Vision and Pattern Recognition


IEEE International Conference on Computer Vision


IEEE International Conference on Pattern Recognition


IEEE International Conference on Image Processing