電腦視覺 Computer and Robot Vision I

Slides:



Advertisements
Similar presentations
Representation and Description
Advertisements

November 12, 2013Computer Vision Lecture 12: Texture 1Signature Another popular method of representing shape is called the signature. In order to compute.
電腦視覺 Computer and Robot Vision I Chapter2: Binary Machine Vision: Thresholding and Segmentation Instructor: Shih-Shinh Huang 1.
Computer Vision Lecture 16: Region Representation
Each pixel is 0 or 1, background or foreground Image processing to
September 10, 2013Computer Vision Lecture 3: Binary Image Processing 1Thresholding Here, the right image is created from the left image by thresholding,
Objective of Computer Vision
Objective of Computer Vision
Lectures 10&11: Representation and description
Information that lets you recognise a region.
Introduction --Classification Shape ContourRegion Structural Syntactic Graph Tree Model-driven Data-driven Perimeter Compactness Eccentricity.
October 8, 2013Computer Vision Lecture 11: The Hough Transform 1 Fitting Curve Models to Edges Most contours can be well described by combining several.
Digital Image Processing
Digital Image Processing
Machine Vision for Robots
8D040 Basis beeldverwerking Feature Extraction Anna Vilanova i Bartrolí Biomedical Image Analysis Group bmia.bmt.tue.nl.
October 14, 2014Computer Vision Lecture 11: Image Segmentation I 1Contours How should we represent contours? A good contour representation should meet.
OBJECT RECOGNITION. The next step in Robot Vision is the Object Recognition. This problem is accomplished using the extracted feature information. The.
Digital Image Processing, 2nd ed. © 2002 R. C. Gonzalez & R. E. Woods Chapter 11 Representation & Description Chapter 11 Representation.
Digital Image Processing Lecture 20: Representation & Description
Recap CSC508.
ENT 273 Object Recognition and Feature Detection Hema C.R.
CS 376b Introduction to Computer Vision 02 / 22 / 2008 Instructor: Michael Eckmann.
Intelligent Vision Systems ENT 496 Object Shape Identification and Representation Hema C.R. Lecture 7.
Digital Image Processing CSC331
September 23, 2014Computer Vision Lecture 5: Binary Image Processing 1 Binary Images Binary images are grayscale images with only two possible levels of.
Digital Image Processing CCS331 Relationships of Pixel 1.
Digital Camera and Computer Vision Laboratory Department of Computer Science and Information Engineering National Taiwan University, Taipei, Taiwan, R.O.C.
Course 9 Texture. Definition: Texture is repeating patterns of local variations in image intensity, which is too fine to be distinguished. Texture evokes.
Digital Image Processing, 2nd ed. © 2002 R. C. Gonzalez & R. E. Woods Representation & Description.
Digital Camera and Computer Vision Laboratory Department of Computer Science and Information Engineering National Taiwan University, Taipei, Taiwan, R.O.C.
Introduction --Classification Shape ContourRegion Structural Syntactic Graph Tree Model-driven Data-driven Perimeter Compactness Eccentricity.
A survey of different shape analysis techniques 1 A Survey of Different Shape Analysis Techniques -- Huang Nan.
Levels of Image Data Representation 4.2. Traditional Image Data Structures 4.3. Hierarchical Data Structures Chapter 4 – Data structures for.
Digital Camera and Computer Vision Laboratory Department of Computer Science and Information Engineering National Taiwan University, Taipei, Taiwan, R.O.C.
CS654: Digital Image Analysis Lecture 36: Feature Extraction and Analysis.
Cartographic Objects Digital Images of a Map Vector Data Model Raster Data Model.
Feature extraction: Corners and blobs. Why extract features? Motivation: panorama stitching We have two images – how do we combine them?
CSSE463: Image Recognition Day 9 Lab 3 (edges) due Weds, 3:25 pm Lab 3 (edges) due Weds, 3:25 pm Take home quiz due Friday, 4:00 pm. Take home quiz due.
1 Overview representing region in 2 ways in terms of its external characteristics (its boundary)  focus on shape characteristics in terms of its internal.
Machine Vision Edge Detection Techniques ENT 273 Lecture 6 Hema C.R.
Robotics Chapter 6 – Machine Vision Dr. Amit Goradia.
Digital Camera and Computer Vision Laboratory Department of Computer Science and Information Engineering National Taiwan University, Taipei, Taiwan, R.O.C.
Digital Image Processing, 3rd ed. © 1992–2008 R. C. Gonzalez & R. E. Woods Gonzalez & Woods Chapter 11 Representation and.
Course 3 Binary Image Binary Images have only two gray levels: “1” and “0”, i.e., black / white. —— save memory —— fast processing —— many features of.
 After an image has been segmented into regions by methods such as those discussed in image segmentation chapter, the segmented pixels usually are represented.
Digital Image Processing CCS331 Relationships of Pixel 1.
Sheng-Fang Huang Chapter 11 part I.  After the image is segmented into regions, how to represent and describe these regions? ◦ In terms of its external.
Materi 09 Analisis Citra dan Visi Komputer Representasi and Deskripsi 1.
Digital Camera and Computer Vision Laboratory Department of Computer Science and Information Engineering National Taiwan University, Taipei, Taiwan, R.O.C.
Image Representation and Description – Representation Schemes
图像处理技术讲座(3) Digital Image Processing (3) Basic Image Operations
IT472: Digital Image Processing
Computer and Robot Vision I
Digital Image Processing Lecture 20: Representation & Description
Materi 10 Analisis Citra dan Visi Komputer
Mean Shift Segmentation
Slope and Curvature Density Functions
Computer Vision Lecture 5: Binary Image Processing
Fitting Curve Models to Edges
Computer Vision Lecture 16: Texture II
Computer and Robot Vision I
Computer and Robot Vision I
CSSE463: Image Recognition Day 9
CSSE463: Image Recognition Day 9
Representation and Description
Fourier Transform of Boundaries
Computer and Robot Vision I
Computer and Robot Vision I
Introduction to Artificial Intelligence Lecture 22: Computer Vision II
Presentation transcript:

電腦視覺 Computer and Robot Vision I Chapter3 Binary Machine Vision: Region Analysis Instructor: Shih-Shinh Huang

Contour-Based Shape Representation Contents Region Properties Simple Global Properties Extremal Points Spatial Moments Mixed Spatial Gray Level Moments Signature Properties Contour-Based Shape Representation

Computer and Robot Vision I Introduction Region Properties

Region Properties Introduction Region Description Region is a segment produced by connected component labeling or signature segmentation. The computation of region properties can be the input for further classification. Gray-Level Value Analysis Shape Property Analysis

Simple Global Properties Region Properties Simple Global Properties Region Area Centroid A=21 r=3.476 c=4.095 1 2 3 4 5 6 7

Simple Global Properties Region Properties Simple Global Properties Perimeter Description It is a sequence of its interior border pixels. Border pixels are the pixels that have some neighboring pixel outside the region. Types of Perimeter 4-Connected Perimeter : Use 8-Connectivity to determine the border pixel. 8-Connected Perimeter :Use 4-Connectivity to determine the border pixel.

Simple Global Properties Region Properties Simple Global Properties 4-Connected Perimeter

Simple Global Properties Region Properties Simple Global Properties 8-Connected Perimeter

Simple Global Properties Region Properties Simple Global Properties Perimeter Representation It is a sequences of border pixels in or are neighborhood

Simple Global Properties Region Properties Simple Global Properties Perimeter Length Vertical or Horizontal Line Diagonal Line

Simple Global Properties Region Properties Simple Global Properties Compactness Measure It is used as a measure of a shape’s compactness. Its smallest value is not for the digital circularity, but it would for continuous planar shapes Octagons Diamonds

Simple Global Properties Region Properties Simple Global Properties Circularity Measure Boundary Pixels

Simple Global Properties Region Properties Simple Global Properties Circularity Measure Properties Digital shape  circular, increases monotonically. It is similar for similar digital/continuous shapes It is orientation and area independent. Polygon Side Estimation

Simple Global Properties Region Properties Simple Global Properties Gray-Level Mean Gray-Level Variance Right hand equation lets us compute variance with only one pass

Simple Global Properties Region Properties Simple Global Properties Microtexture Properties Co-occurrence Matrix S : a set of all pairs of pixels that are in some defined spatial relationship (4-neighbors)

Simple Global Properties Region Properties Simple Global Properties 0 1 2 3 1 2 3 DC & CV Lab. CSIE NTU

Simple Global Properties Region Properties Simple Global Properties Microtexture Properties Texture Second Moment Texture Entropy Texture Homogeneity

Simple Global Properties Region Properties Simple Global Properties Microtexture Properties Contrast Correlation

Definition of Extremal Points Region Properties Extremal Points Definition of Extremal Points It has an extremal coordinate value in either its row or column coordinate position They can be as many as eight distinct extermal points.

Region Properties Extremal Points

Different extremal points may be coincident Region Properties Extremal Points Different extremal points may be coincident

Definition of Extremal Coordinate Region Properties Extremal Points Definition of Extremal Coordinate Topmost Bottommost Leftmost Rightmost Topmost Rightmost Leftmost Bottommost

Definition of Extremal Coordinate Region Properties Extremal Points Definition of Extremal Coordinate Topmost Left Topmost Right Topmost Right Topmost Left

Respective Axes (M1, M2, M3, M4) Region Properties Extremal Points Respective Axes (M1, M2, M3, M4) Form by each pair of opposite extremal points M1: Topmost Left & Bottommost Right M2: Topmost Right & Bottommost Left M3: Rightmost Top & Leftmost Bottom M4: Rightmost Bottom& Leftmost Top. Properties Length Orientation

Region Properties Extremal Points

Length of Respective Axes Region Properties Extremal Points Length of Respective Axes : one end point of respective axes : the other point of respective axes Quantization Error Compensation Term

Orientation of Respective Axes Region Properties Extremal Points Orientation of Respective Axes Orientation of a line segment is taken as counterclockwise with respect to column axis. Quantization Error Compensation Term

Properties of Line-like Region Region Properties Extremal Points Properties of Line-like Region Major Axis : the axis with the largest length. The length and orientation of major axis stands for the same thing for this region.

Region Properties Extremal Points Properties of Line-like Region

Properties of Triangular Shapes Region Properties Extremal Points Properties of Triangular Shapes Apex Selection: Find the extremal point having the greatest sum of its two largest distances. Extremal Point Distance Objective Function

Properties of Triangular Shapes Region Properties Extremal Points Properties of Triangular Shapes Side Length Base Length Altitude Height

Region Properties Extremal Points

Second-Order Spatial Moments Region Properties Spatial Moments Second-Order Spatial Moments Row Moment Mixed Moment Column Moment

Second-Order Spatial Moments Region Properties Spatial Moments Second-Order Spatial Moments They have value meaning for a region of any shape Similarly, the covariance matrix has value and meaning for any two-dimensional pdf. Example: An ellipse A whose center is the origin.

Mixed Spatial Gray Level Moments Region Properties Mixed Spatial Gray Level Moments Description A property that mixes up two properties. Spatial Properties: Region Shape, Position Intensity properties Two Second-order Mixed Spatial Gray Properties

Mixed Spatial Gray Level Moments Region Properties Mixed Spatial Gray Level Moments Application: Determine the least-square, best-fit gray level intensity plane. Unknowns Variables: Objective Function

Mixed Spatial Gray Level Moments Region Properties Mixed Spatial Gray Level Moments Application: Determine the least-square, best-fit gray level intensity plane Take partial derivative of with respect to Least Square Method

Mixed Spatial Gray Level Moments Region Properties Mixed Spatial Gray Level Moments Application: Determine the least-square, best-fit gray level intensity plane

Mixed Spatial Gray Level Moments Region Properties Mixed Spatial Gray Level Moments

Computer and Robot Vision I Introduction Signature Properties

Signature Properties Introduction Signature Review Remark: Signature analysis is important because of easy, fast implementation in pipeline hardware

Signature Computation Signature Properties Signature Computation Centroid Second-Order Moment

Signature Computation Signature Properties Signature Computation Second-Order Moment

Circle Center Determination Signature Properties Circle Center Determination Description We can determine the center position of circular region from signature analysis.

Circle Center Determination Signature Properties Circle Center Determination Derivation

Circle Center Determination Signature Properties Circle Center Determination Derivation Compute by a table-look-up technique

Circle Center Determination Algorithm Step 1: Partition the circuit into four quadrants formed by two orthogonal lines intersecting inside the circle. Step 2: Using signature analysis to compute the areas A, B, C, and, D. Step 3: Compute using the derived equation.

Computer and Robot Vision I Introduction Contour-Based Shape Representation

Chain Code Description It describes an object by a sequence of unit-size line segment with a given orientation. The first element must bear information about its position to permit region reconstruction. Chain Code: 3, 0, 0, 3, 0, 1, 1, 2, 1, 2, 3, 2

Chain Code Matching Requirement It must be independent of the choice of the first border pixel in the sequence. It requires the normalization of chain code Interpret the chain code as a base 4 number. Find the pixel in the border sequence which results in the minimum integer number. Chain Code: (300301121232)4 Chain Code: (003011212323)4

Curvature Description Curvature is defined as the rate of change of slope in the continuous case. The evaluation algorithm in the discrete case is based on the detection of angles between two lines. Values of the curvature at all boundary pixels can be represented by a histogram for matching.

Curvature b: sensitivity to local changes.

Signature Description The signature is a sequence of normal contour distances. It can be calculated for each boundary elements as a function of the path length.

Definition of Chord Distribution Description Chord is a line joining any two points of the region boundary. The distribution of lengths and angles of all chords may be used for shape description. Definition of Chord Distribution : contour points : all other points

Chord Distribution Rotation-Independent Radial Distribution

Recursive Boundary Splitting Segment Sequence Description It is a way to represent the boundary using segments with specified properties. Recursive Boundary Splitting

Chromosomes Representation. Segment Sequence Structure Description Curves are segmented into several types Circular Arcs Straight Line Segments are considered as primitives for syntactic shape recognition Chromosomes Representation.

Scale-Space Image Description Approach Properties Sensitivity of shape descriptors to scale (image resolution) is an undesirable feature. Some curve segmentation points exist in one resolution and disappear in others. Approach Properties Only new segmentation points can appear at higher resolution. No existing segmentation points can disappear.

Scale-Space Image Approach Description It is based on application of a unique Gaussian smoothing kernel to a one-dimensional signal. The zero-crossing of the second derivative is detected to determine the peak of curvature. The positions of zero-crossing give the positions of curve segmentation points.

Scale-Space Image

Computer and Robot Vision I The End