object recognition. the next step in robot vision is the object recognition. this problem is...

61
OBJECT RECOGNITION

Upload: dortha-dixon

Post on 29-Dec-2015

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

OBJECT RECOGNITION

Page 2: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

OBJECT RECOGNITION

The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information.

The object recognition algorithm is to be powerful and fast so that the required object is uniquely recognized.

Page 3: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Image Features

Image feature refers either to

– a global property of an image or a part of it– for example: average grey level, area in pixels – recognition properties important

or

– part of image with some special properties– for example: line, circle, textured region– location important

Page 4: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Good Features

● Meaningful– associated to interesting scene elements typically sharp intensity variations (contours of objects), or– regions of uniform intensity (planar surfaces)

● Detectable– we must be able to locate the features algorithmically– different algorithms for different features

● We can not get perfect features

Page 5: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Feature types

• Geometrical entities

– points• for example: edge points, corner points, interest

points

• regions

• lines

• circles, ellipses

• contours

Page 6: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Edge Detection – Problem

Page 7: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Edges

• Edge points are pixels where image values undergo a sharp variation

Page 8: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Corner Detection

• Corners are well defined in location

– useful features to track in image sequences

– should be understood as corner structures in

image intensity, not corners of objects

● Image gradient tells us about changes

– How to calculate the gradient?

– How does gradient behave near corners?

Page 9: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

There are mainly two methods for object recognition

(a)Template matching technique

A template is provided to computer and the computer is trained to match the object with the template irrespective of object orientation.

Page 10: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

(b)Structural technique

Several structural techniques are available. These may take the features also in to account. We discuss a common method known as

CHAIN CODE. Here there are two approaches

(a) 4-Directional Chain Code

(b) 8 Directional Chain code

We shall describe (a) first; (b) is an extension of (a).

Each object is coded in terms of this 4-directional chain code

Page 11: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

STARTStart with the start point and go along the arrows

Comparing the contour of object with respect to 4-directional chain code converter, we get

Chain code of Object: 03032211

4-directional converter

8-directional converter

0

1

2

3

0

12

3

4

5

67

OBJECT

Page 12: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

CHAIN CODE : 03032211

We then get the difference;

Difference between 0 and 3 is 3 (ref 4 direc. converter)

Difference between 3 and 0 is 1

Difference between 0 and 3 is 3 and so on

Difference between 1 and 1 is 0

Difference between 1 and 0 is 3.

Difference Code is : 31330303

Take the minimum (decimal)value of Difference code as

SHAPE NUMBER: 03033133

This SHAPE number is for the object, uniquely recognized, independent of rotation (by 900)

Page 13: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Another Object

Verify:

Chain Code:

0330011033323333221112111

Difference code:

3010103300310003030013003

Shape Number:

0003030013003301010330031

Page 14: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

If the object edges are of slopes 450, 1350 , - 450 and - 1350, then we can use the 8-Directional converter. The procedure of getting the shape number is the same.

This method of object recognition is fast and can be used for different shapes of objects to be recognized if they are coming in a random sequence.

Page 15: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Advantage of Chain Coding

• Reduce storage space as number and value of pixels are not required.

• It can be use to compute the values of some features of objects such as area, perimeter, width etc

Page 16: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

FEATURE EXTRACTION:In robot vision, it is often necessary to

distinguish one object from another. This is accomplished by mean of features that uniquely characterize the object.

Some features of objects that can be used in Robot Vision are:

(a) Area (b) Minimum Enclosing rectangle

(c) Diameter (d) center of gravity

(e)Perimeter (f) eccentricity (g) Aspect Ratio

(h) Number of holes (i) Moments

We shall illustrate the features through an example:

Page 17: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• Image Area

• Center of Gravity (COG)

Or centroid for xc and yc. The area and COG is used to identify the position of the object

m

x

n

y

yxIA1 1

),(

m

x

n

yc yxxIA

x1 1

),(1

m

x

n

yc yxyIA

y1 1

),(1

Page 18: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• Moments- A sequence of numbers characterzing the shape of an object

• The sum of power (j+k) is the order of the moment

0,0),(,

kjyxIyxMOyx

kjjk

m

x

n

y

yxIAM1 1

00 ),(

m

x

n

yc yxxIA

xM1 1

10 ),(1

m

x

n

yc yxyIA

yM1 1

01 ),(1

Page 19: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• If the COG is known we can determine the central of moment

• Because object is balanced at COG, the first order moment is zero

0,0),()()(,

kjyxIyyxxOyx

kc

jcjk

m

x

n

yc

m

x

n

yc

yxIyy

yxIxx

1 101

1 110

0),()(

0),()(

Page 20: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• The second order moment give the moment inertia of the image

m

x

n

ycc

m

x

n

yc

m

x

n

yc

yxIyyxx

yxIyy

yxIxx

1 111

1 1

202

1 1

220

),())((

),()(

),()(

Page 21: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• Orientation-the angle of inclination ),2(2tan

2

1022011 A

Page 22: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• Eccentricity- maximum chord length is along the principal axis or major axis of object and minimum chord length is perpendicular to major axis

lengthchordMinimum

lengthchordMaximumtyEccentrici

22

,)()(max ijij

jiyyxx

lengthchordMaximum

Page 23: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• Roundness,

• Aspect Ratio=Length of Rectangle enclosing object» Width of rectangle enclosing Object

A

yxR cc

22

Page 24: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

EXAMPLE: Let an original image of an object undergone several image processing techniques and finally available to us as a pixel pattern shown below:

Page 25: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Some of features can be computed as:

(a) Moment ( M00 ) = = 24

(b)Eccentricity = (Max x-length) / (Max y-length) = 9/4

(c) Perimeter = 26

(d)Area = 24

(e) Diameter = 9

(f) Thinness = {Diameter / area }

= ( 9 / 24 ) = 0.375

(g) compactness =

{ (perimeter)2 / area } = ( 26 2 /24 ) = 28.17

(a) No: of holes = 0

Objects having these features belong to one category

yxyx

,

00

Page 26: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Example

Page 27: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 28: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 29: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Matching

• Recognition techniques based on matching represent each class by a proto-type pattern vector.

• An unknown pattern is assigned to the class to which it is closest in terms of a predetermine metric.

• The common approach is by minimum distance classifier

Page 30: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Some of the edge identified products

Page 31: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Minimum distance classifier

• In image processing system, an object is described by a pattern vector

• Pattern recognition is to develop or use some techniques or algorithms to recognize objects in images.

• Minimum distance classifier is commonly used and based decision functions.

Page 32: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Decision Function

• The function is to determine decision boundary to separate different pattern classes.

Page 33: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 34: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 35: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 36: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 37: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 38: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 39: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 40: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 41: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• Example:

Page 42: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 43: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 44: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• Solution:

Page 45: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 46: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 47: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 48: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 49: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 50: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 51: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Optimum Statistical Classifier

• Probability consideration become important in pattern recognition under random pattern classes

Page 52: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 53: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 54: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 55: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 56: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

• For the above equation to hold we must know the probability function.

• Thus, the Gaussian probability density function is considered

Page 57: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The

Bayes Classifier for Gausssian pattern classes

Page 58: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 59: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 60: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The
Page 61: OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The