Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.
-
Upload
merry-nicholson -
Category
Documents
-
view
223 -
download
1
Transcript of Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.
![Page 1: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/1.jpg)
Shape-Representationand
Shape SimilarityCIS 601
by Rolf Lakaemper
modified by Longin Jan Latecki
![Page 2: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/2.jpg)
Motivation
WHY SHAPE ?
![Page 3: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/3.jpg)
Motivation
We’ve seen this already in the introductionof this course:These objects are recognized by…
![Page 4: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/4.jpg)
Motivation
These objects are recognized by…
Texture Color Context Shape
X X
X X
X
X
X
X X
![Page 5: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/5.jpg)
Motivation
Shape is not the only, but a verypowerful descriptor of
image content
![Page 6: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/6.jpg)
Why Shape ?
Several applications in computer vision use shape processing:
• Object recognition• Image retrieval
• Processing of pictorial information• Video compression (eg. MPEG-7)
…
![Page 7: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/7.jpg)
ISS Database
Example 2: ISS-Databasehttp://knight.cis.temple.edu/~shape
![Page 8: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/8.jpg)
The Interface (JAVA – Applet)
![Page 9: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/9.jpg)
ISS Database
ISS: Query by Shape
Sketch of Shape
Query:
by Shape only
Result:
Satisfying ?
![Page 10: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/10.jpg)
ISS Database
The ISS-Database will be topicof this tutorial
![Page 11: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/11.jpg)
Overview
Overview
• Why shape ?• What is shape ?• Shape similarity• (Metrices)
• Classes of similarity measures• (Feature Based Coding)
• Examples for global similarity
![Page 12: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/12.jpg)
Why Shape ?
Why Shape ?
• Shape is probably the most important property that is perceived about objects. It allows to predict more facts about an object than other features, e.g. color (Palmer 1999)
• Thus, recognizing shape is crucial for object recognition. In some applications it may be the only feature present, e.g. logo recognition
![Page 13: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/13.jpg)
Why Shape ?
Shape is not only perceived by visual means:
• tactical sensors can also provide shape information that are processed in a similar way.
• robots’ range sensor provide shape information, too.
![Page 14: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/14.jpg)
Shape
Typical problems:
• How to describe shape ? • What is the matching
transformation?• No one-to-one correspondence• Occlusion• Noise
![Page 15: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/15.jpg)
Shape
• Partial match: only a part of the query appears in a part of the database shape
![Page 16: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/16.jpg)
What is Shape ?
What is Shape ?
Plato, "Meno", 380 BC:
• "figure is the only existing thing that is found always following color“
• "figure is limit of solid"
![Page 17: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/17.jpg)
What is Shape ?
… let’s start with some properties easier to agree on:
• Shape describes a spatial regionShape is a (the ?) specific part of spatial cognition
• Typically addresses 2D space
![Page 18: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/18.jpg)
What is Shape ?
• 3D => 2D projection
![Page 19: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/19.jpg)
What is Shape ?
• the original 3D (?) object
![Page 20: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/20.jpg)
What is Shape ?
Moving on from the naive understanding, some questions arise:
• Is there a maximum size for a shape to be a shape?
• Can a shape have holes?• Does shape always describe a connected
region?• How to deal with/represent partial shapes
(occlusion / partial match) ?
![Page 21: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/21.jpg)
What is Shape ?
Shape or Not ?
Continuous transformation from shape to two shapes: Is there a point when it stops being a single shape?
![Page 22: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/22.jpg)
What is Shape ?
But there’s no doubt that
a single, connected region
is a shape.
Right ?
![Page 23: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/23.jpg)
What is Shape ?
A single, connected region.But a shape ?
A question of scale !
![Page 24: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/24.jpg)
What is Shape ?
• There’s no easy, single definition of shape• In difference to geometry, arbitrary shape is not
covered by an axiomatic system
• Different applications in object recognition focus on different shape related features• Special shapes can be handled
• Typically, applications in object recognition employ a similarity measure to determine a plausibility that two shapes correspond to each other
![Page 25: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/25.jpg)
Similarity
So the new question is:
What is Shape Similarity ?
or
How to Define a Similarity Measure
![Page 26: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/26.jpg)
Similarity
Again: it’s not so simple (sorry).
There’s nothing like
THE
similarity measure
![Page 27: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/27.jpg)
Similarity
which similarity measure,
depends onwhich required properties,
depends onwhich particular matching problem,
depends onwhich application
![Page 28: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/28.jpg)
Similarity
... robustness
... invariance to basic transformations
Simple Recognition (yes / no)
Common Rating (best of ...)
Analytical Rating (best of, but...)
…which application
![Page 29: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/29.jpg)
Similarity
…which problem
• computation problem: d(A,B)
• decision problem: d(A,B) <e ?
• decision problem: is there g: d(g(A),B) <e ?
• optimization problem: find g: min d(g(A),B)
![Page 30: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/30.jpg)
Similarity
…which properties:
We concentrate here on the computational problem d(A,B)
![Page 31: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/31.jpg)
Similarity Measure
Requirements to a similarity measure
• Should not incorporate context knowledge (no AI), thus computes generic shape similarity
![Page 32: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/32.jpg)
Similarity Measure
Requirements to a similarity measure
• Must be able to deal with noise• Must be invariant with respect to basic
transformations
Next:StrategyScaling (or resolution)
Rotation
Rigid / non-rigid deformation
![Page 33: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/33.jpg)
Similarity Measure
Requirements to a similarity measure
• Must be able to deal with noise
• Must be invariant with respect to basic transformations
• Must be in accord with human perception
![Page 34: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/34.jpg)
Similarity Measure
Some other aspects worth consideration:
• Similarity of structure• Similarity of area
Can all these aspects be expressed by a single number?
![Page 35: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/35.jpg)
Similarity Measure
Desired Properties of a Similarity Function C(Basri et al. 1998)
• C should be a metric• C should be continuous• C should be invariant (to…)
![Page 36: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/36.jpg)
Properties
Metric Properties
S set of patternsMetric: d: S S R satisfying1. Self-identity: xS, d(x,x)=02. Positivity: x yS, d(x,y)>03. Symmetry: x, yS, d(x,y)= d(y,x)4. Triangle inequality: x, y, zS, d(x,z)d(x,y)
+d(y,z)
• Semi-metric: 1, 2, 3• Pseudo-metric: 1, 3, 4• S with fixed metric d is called metric space
![Page 37: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/37.jpg)
Properties
1. Self-identity: xS, d(x,x)=02. Positivity: x yS, d(x,y)>0
…surely makes sense
![Page 38: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/38.jpg)
Properties
![Page 39: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/39.jpg)
Properties
![Page 40: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/40.jpg)
Properties
In general:
• a similarity measure in accordance with human perception is NOT a metric. This leads to deep problems in further processing, e.g. clustering, since most of these algorithms need metric spaces !
![Page 41: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/41.jpg)
Properties
![Page 42: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/42.jpg)
Properties
![Page 43: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/43.jpg)
Properties
Some more properties:
• One major difference should cause a greater dissimilarity than some minor ones.
• S must not diverge for curves that are not smooth (e.g. polygons).
![Page 44: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/44.jpg)
Similarity Measures
Classes of Similarity Measures:
Similarity Measure depends on
• Shape Representation
• Boundary
• Area (discrete: = point set)
• Structural (e.g. Skeleton)
• Comparison Model
• feature vector
• direct
![Page 45: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/45.jpg)
Similarity Measures
direct feature based
Boundary Spring model, Cum. Angular Function, Chaincode, Arc Decomposition (ASR-Algorithm)
Central Dist. Fourier
Distance histogram
…
Area (point set) Hausdorff
…
Moments
Zernike Moments
…
Structure Skeleton
…
---
![Page 46: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/46.jpg)
Feature Based Coding
Feature Based Coding (again…)
This category defines all approaches that determine a feature-vector for a given shape.
Two operations need to be defined: a mapping of shape into the feature space and a similarity of feature vectors.
Representation Feature Extraction Vector Comparison
![Page 47: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/47.jpg)
Vector Comparison
Another feature you should have heard of:
(Discrete) MomentsShape A,B given as
• Area (continuous) or
• Point Sets (discrete)
![Page 48: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/48.jpg)
Moments
Discrete Point Sets
![Page 49: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/49.jpg)
Moments
![Page 50: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/50.jpg)
Moments
![Page 51: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/51.jpg)
Moments
![Page 52: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/52.jpg)
Discrete Moments
Exercise:
Please compute all 7 moments for the following shapes, compare the vectors using different comparison techniques
![Page 53: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/53.jpg)
Discrete Moments
Result: each shape is transformed to a 7-dimensional vector. To compare the shapes, compare the vectors (how ?).
![Page 54: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/54.jpg)
3D Distance Histogram
Another Example
3D Distance Histogram
Shape A,B given as 3D point set
![Page 55: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/55.jpg)
3D Distance Histogram
![Page 56: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/56.jpg)
Feature Based Coding
Again:
Two operations need to be defined: a mapping of shape into the feature space and a similarity of feature vectors.
We hence have TWO TIMES an information reduction of the basic representation, which by itself is already a mapping of the ‘reality’.
Representation Feature Extraction Vector Comparison
![Page 57: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/57.jpg)
Direct Comparison
End of Feature Based Coding !
Next:
Direct Comparison
![Page 58: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/58.jpg)
Vector ComparisonDirect Comparison
Example 1
Hausdorff Distance
Shape A,B given as point sets
A={a1,a2,…}
B={b1,b2,…}
![Page 59: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/59.jpg)
Vector ComparisonFeature Based Coding
![Page 60: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/60.jpg)
Vector ComparisonHausdorff Distance
![Page 61: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/61.jpg)
Vector ComparisonHausdorff Distance
Hausdorff:
Unstable with respect to noise(This is easy to fix ! How ?)
Problem: Invariance !Nevertheless: Hausdorff is the motor behind many applications in specific fields (e.g. character recognition)
![Page 62: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/62.jpg)
Vector ComparisonBoundary Representation
Example 2
Chain code Comparison
Shape A,B given as chain codes
![Page 63: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/63.jpg)
Vector ComparisonBoundary Representation
![Page 64: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/64.jpg)
Getting Boundaries
As output of image segmentation, we obtain objects that can be viewed as bitmaps.
Let f be a bitmap, i.e., a binary image with 0s representing the background. We can obtain the boundary of the object represented with f using Matlab function:B = boundaries(f);the obtained boundary is 8-connected.
![Page 65: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/65.jpg)
I = imread('pout.tif');figure, imshow(I);figure, imhist(I);BW = im2bw(I, 0.45); % makes a binary image
% all pixels above 0.4*255 are 1 and % the rest is 0 which is black
figure; imshow(BW) ;
B = boundaries(BW);figure; imshow(B);
![Page 66: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/66.jpg)
Homework 10 For certain images, objects of interest can be
segmented using simple tools.Your task: compute the rabbit's boundary.
Link to the image.
![Page 67: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/67.jpg)
The original image. All pixels having a greater red than green ratio After two stages of morphological processing,
we are ready to get the contour.
![Page 68: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/68.jpg)
Vector ComparisonBoundary Representation
A binary image can be converted into a ‘chain code’ representing the boundary. The boundary is traversed and a string representing the curvature is constructed.
0
123
4
5 6 7
C
5,6,6,3,3,4,3,2,3,4,5,3,…
![Page 69: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/69.jpg)
Chain Code
![Page 70: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/70.jpg)
![Page 71: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/71.jpg)
![Page 72: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/72.jpg)
Vector ComparisonBoundary Representation
Resulting strings are then compared using classical string-matching techniques.
Not very robust.
![Page 73: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/73.jpg)
Vector ComparisonBoundary Representation
Digital curves suffer from effects caused by digitalization, e.g. rotation:
![Page 74: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/74.jpg)
Shape Signatures
[st, angle, x0,y0] = signature(B, x0, y0);figure; plot(angle, st);
Resulting strings are then compared using string-matching techniques.
![Page 75: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/75.jpg)
![Page 76: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/76.jpg)
Vector ComparisonStructural Representation
Structural approaches capture the
structure of a shape, typically by
representing shape as a graph.
Typical example: skeletons
![Page 77: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/77.jpg)
Vector ComparisonStructural Representation
Skeletons
Shape A,B primarily given as area or boundary, structure is derived from
representation
![Page 78: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/78.jpg)
Vector ComparisonStructural Representation
The computation can be described as a medial axis transform, a kind of discrete generalized voronoi.
![Page 79: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/79.jpg)
Vector ComparisonStructural Representation
The graph is constructed mirroring the adjacency of the skeleton’s parts. Edges are labeled according to the qualitative classes.
Matching two shapes requires matching two usually different graphs against each other.
![Page 80: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/80.jpg)
Vector ComparisonStructural Representation
Problems of skeletons:
- Pruning
![Page 81: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/81.jpg)
Vector ComparisonStructural Representation
-Robustness
![Page 82: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/82.jpg)
Vector ComparisonShape similarity
All similarity measures shown can not deal with occlusions or partial matching (except skeletons ?) !
They are useful (and used) for specific applications, but are not sufficient to deal with arbitrary shapes
Solution: Part – based similarity !
![Page 83: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/83.jpg)
Shape-Representation
and
Shape SimilarityPART 2: PART BASED SIMILARITY
![Page 84: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/84.jpg)
Motivation
WHY PARTS ?
![Page 85: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/85.jpg)
Motivation
![Page 86: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/86.jpg)
Motivation
![Page 87: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/87.jpg)
Motivation
![Page 88: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/88.jpg)
Motivation
Global similarity measures fail at:
• Occlusion• Global Deformation• Partial Match• (actually everything that occurs under ‘real’ conditions)
![Page 89: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/89.jpg)
Parts
Requirements for a Part Based Shape Representation
(Siddiqi / Kimia ’96: ‘Parts of Visual Form: Computational Aspects’)
![Page 90: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/90.jpg)
Parts
How should parts be defined / computed ?
Some approaches:
• Decomposition of interior• Skeletons• Maximally convex parts• Best combination of primitives
• Boundary Based• High Curvature Points• Constant Curvature Segments
![Page 91: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/91.jpg)
Parts
Principal approach:
Hoffman/Richards (’85):
‘Part decomposition should precede part description’
=> No primitives, but general principles
![Page 92: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/92.jpg)
Parts
No primitives, but general principals
“When two arbitrarily shaped surfaces are made to interpenetrate they always meet in a contour of concave discontinuity of their tangent planes” (transversality principle)
![Page 93: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/93.jpg)
Parts
“When two arbitrarily shaped surfaces are made to interpenetrate they always meet in a contour of concave discontinuity of their tangent planes” (transversality principle)
Divide a plane curve into parts at negative minima of curvature
![Page 94: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/94.jpg)
Parts
Different notions of parts:
• Parts: object is composed of rigid parts
• Protrusions: object arises from object by deformation due to a (growth) process (morphology)
• Bends: Parts are result of bending the base object
![Page 95: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/95.jpg)
Parts
The Shape Triangle
![Page 96: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/96.jpg)
Parts
This lecture focuses on parts, i.e. on partitioning a shape
![Page 97: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/97.jpg)
Framework
A Framework for a Partitioning Scheme
Scheme must be invariant to 2 classes of changes:
• Global changes : translations, rotations & scaling of 2D shape, viewpoint,…
• Local changes: occlusions, movement of parts (rigid/non-rigid deformation)
![Page 98: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/98.jpg)
Framework
A general decomposition of a shape should be based on the
interaction between two parts rather than on their shapes.
-> Partitioning by Part Lines
![Page 99: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/99.jpg)
Framework
Definition 1:
A part line is a curve whose end points rest on the boundary of the shape, which is entirely embedded in it, and which divides it into two connected components.
![Page 100: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/100.jpg)
Definition 2:
A partitioning scheme is a mapping of a connected region in the image to a finite set of connected regions separated by part-lines.
Framework
![Page 101: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/101.jpg)
Definition 3:
A partitioning scheme is invariant if the part lines of a shape that is transformed by a combination of translations, rotations and scalings are transformed in exactly the same manner.
Framework
![Page 102: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/102.jpg)
Definition 4:
A partitioning scheme is robust if for any two shapes A and B, which are exactly the same in some neighborhood N, the part lines contained in N for A and B are exactly equivalent.
Framework
![Page 103: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/103.jpg)
Definition 5:
A partitioning scheme is stable if slight deformations of the boundary of a shape cause only slight changes in its part lines
Framework
![Page 104: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/104.jpg)
Definition 6:
A partitioning scheme is scale-tuned if when moving from coarse to fine scale, part lines are only added, not removed, leading to a hierarchy of parts.
Framework
![Page 105: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/105.jpg)
A general purpose partitioning scheme that is consistent with
these requirements is the partitioning by
limbs and necks
Framework
![Page 106: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/106.jpg)
Definition :
A limb is a part-line going through a pair of negative curvature minima with co-circular boundary tangents on (at least) one side of the part-line
Limbs and Necks
![Page 107: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/107.jpg)
Motivation: co-circularity
Limbs and Necks
The decomposition of the right figure is no longer intuitive: absence of ‘good continuation’
![Page 108: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/108.jpg)
Smooth continuation: an example for
form from function
• Shape of object is given by natural function
• Different parts having different functions show sharp changes in the 3d surface of the connection
• Projection to 2d yields high curvature points
Limbs and Necks
![Page 109: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/109.jpg)
Examples of limb based parts
Limbs and Necks
![Page 110: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/110.jpg)
Definition :A neck is a part-line which is also a
local minimum of the diameter of an inscribed circle
Limbs and Necks
![Page 111: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/111.jpg)
Motivation for necks: Form From Function
• Natural requirements (e.g. space for articulation and economy of mass at the connection) lead to a narrowing of the joint between two parts
Limbs and Necks
![Page 112: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/112.jpg)
The Limb and Neck partitioning scheme is consistent with the
previously defined requirements
• Invariance• Robustness• Stability• Scale tuning
Limbs and Necks
![Page 113: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/113.jpg)
Examples:
Limbs and Necks
![Page 114: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/114.jpg)
The scheme presented does NOT include a similarity measure !
Limbs and Necks
![Page 115: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/115.jpg)
Part Respecting Similarity Measures
Algorithms
![Page 116: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/116.jpg)
Curvature Scale Space(Mokhtarian/Abbasi/Kittler)
A similarity measure implicitely respecting parts
CSS
![Page 117: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/117.jpg)
CSS
Creation of reflection-point based feature-vector which implicitly contains part – information
![Page 118: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/118.jpg)
CSS
Properties:
• Boundary Based• Continuous Model (!) • Computes Feature Vector
• compact representation of shape• Performs well !
![Page 119: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/119.jpg)
CSS
Some results (Database: 450 marine animals)
![Page 120: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/120.jpg)
CSS
The main problem:
CSS is continuous, the computer vision world is discrete.
How to measure curvature in discrete boundaries ?
![Page 121: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/121.jpg)
Dominant Points
Local curvature = average curvature in ‘region of support’
To define regions of support, ‘dominant points’ are needed !
![Page 122: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/122.jpg)
Dominant Points
Dominant Points(“Things should be expressed as simple as possible, but not simpler”,
A. Einstein)
Idea: given a discrete boundary S compute polygonal boundary S’ with minimum number of vertices which is
visually similar to S.
![Page 123: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/123.jpg)
Dominant Points
Example Algorithms( 3 of billions…)
• Ramer• Line Fitting
• Discrete Curve Evolution
![Page 124: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/124.jpg)
DCE
Discrete Curve Evolution(Latecki / Lakaemper ’99)
Idea:
Detect subset of visually significant points
![Page 125: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/125.jpg)
Discrete Curve Evolution (DCE)
We achieve a comparable level of detail with DCE.
Before a similarity measure is applied, the shape of objects is simplified by DCE in order to
• reduce influence of noise,
• simplify the shape by removing irrelevant shape features without changing relevant shape features.
![Page 126: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/126.jpg)
Curve Evolution
Target: reduce data by elimination of irrelevant features, preserve relevant features
... noise reduction
... shape simplification:
![Page 127: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/127.jpg)
Discrete Curve Evolution (DCE)
u
v
w u
v
w
It yields a sequence: P=P0, ..., Pm
Pi+1 is obtained from Pi by deleting the vertices of Pi that have minimal relevance measure
K(v, Pi) = |d(u,v)+d(v,w)-d(u,w)|
>
![Page 128: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/128.jpg)
Discrete Curve Evolution: Preservation of position, no blurring
![Page 129: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/129.jpg)
Discrete Curve Evolution: robustness with respect to noise
![Page 130: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/130.jpg)
Discrete Curve Evolution: extraction of linear segments
![Page 131: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/131.jpg)
Discrete Curve Evolution: mathematical properties
Convexity Theorem (trivial)Discrete curve evolution (when applied to a polygon)
converges to a convex polygon. Continuity Theorem (nontrivial)Discrete curve evolution is continuous.
L. J. Latecki, R.-R. Ghadially, R. Lakämper, and U. Eckhardt: Continuity of the discrete curve evolution. Journal of Electronic Imaging 9, pp. 317-326, 2000.
Polygon Recovery (nontrivial)DCE allows to recover polygons from their digital images.L.J. Latecki and A. Rosenfeld: Recovering a Polygon form Noisy Data.
Computer Vision and Image Understanding (CVIU) 86, 1-20, 2002.
![Page 132: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/132.jpg)
Comparable level of detail for DCE (=stop condition) is based on a threshold on the relevance measure
![Page 133: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/133.jpg)
Comparable level of detail for DCE is based on a threshold on the relevance measure
![Page 134: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/134.jpg)
Scale Space Approaches to Curve Evolution
1. reaction-diffusion PDEs
2. polygonal analogs of the PDE-evolution (Bruckstein et al. 1995)
3. approximation (e.g., Bengtsson and Eklundh 1991)
Main differences:
[to 1, 2:] Each vertex of the polygon is moved at a single evolution step, whereas in our approach the remaining vertices do not change their positions.
[to 1, 3:] Our approach is parameter-free(we only need a stop condition)
![Page 135: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/135.jpg)
![Page 136: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/136.jpg)
The evolution...
... reduces the shape-complexity
... is robust to noise
... is invariant to translation, scaling and rotation
... preserves the position of important vertices
... extracts line segments
... is in accord with visual perception
... offers noise-reduction and shape abstraction
... is parameter free
Curve Evolution: Properties
... is translatable to higher dimensions
![Page 137: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/137.jpg)
Extendable to higher dimensions
Curve Evolution: Properties
![Page 138: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/138.jpg)
Extendable to higher dimensions
Curve Evolution: Properties
![Page 139: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/139.jpg)
Extendable to higher dimensions
Curve Evolution: Properties
![Page 140: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/140.jpg)
Extendable to higher dimensions
Curve Evolution: Properties
![Page 141: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/141.jpg)
Scale Space
Ordered set of representations on different information levels
![Page 142: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/142.jpg)
The polygonal representation achieved by the DCE has a huge
advantage:
It allows easy boundary partitioning using convex / concave
parts (remember the limbs !)
Polygonal Representation
![Page 143: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/143.jpg)
Some results of part line decomposition:
DCE
![Page 144: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/144.jpg)
The ASR (Advanced Shape Recognition) Algorithm uses the boundary parts achieved by the
polygonal representation for a part based similarity measure !
(Note: this is NOT the area partitioning shown in the previous slide)
ASR
![Page 145: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/145.jpg)
Behind The Scenes of the ISS - Database:
Modern Techniques of ShapeRecognition and Database Retrieval
How does it work ?
![Page 146: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/146.jpg)
The 2nd Step First: Shape Comparison
Developed by Hamburg University in cooperation withSiemens AG, Munich, for industrial applications in...
... robotics
... multimedia (MPEG – 7)
ISS implements the ASR (Advanced Shape Recognition) Algorithm
![Page 147: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/147.jpg)
Reticent Proudness…
MPEG-7: ASR outperforms classical approaches !
Similarity test (70 basic shapes, 20 different deformations):
Wavelet Contour Heinrich Hertz Institute Berlin 67.67 %
Multilayer Eigenvector Hyundai 70.33 %
Curvature Scale Space Mitsubishi ITE-VIL 75.44 %
ASR Hamburg Univ./Siemens AG 76.45 %
DAG Ordered Trees Mitsubishi/Princeton University 60.00 %
Zernicke Moments Hanyang University 70.22 %
(Capitulation :-) IBM --.-- %
![Page 148: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/148.jpg)
ASR: StrategyASR: Strategy
Source: 2D - Image
Arc – Matching
Contour – Segmentation
Contour Extraction
Object - Segmentation
Evolution
![Page 149: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/149.jpg)
ASR: StrategyASR: Strategy
Arc – Matching
DCE
Contour – Segmentation
![Page 150: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/150.jpg)
Contour Segmentation
Correspondence ?
Similarity of parts ?
![Page 151: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/151.jpg)
Part Similarity
Similarity of parts ?
= Boundary Similarity Measure
= Similarity of polygons
![Page 152: Shape-Representation and Shape Similarity CIS 601 by Rolf Lakaemper modified by Longin Jan Latecki.](https://reader035.fdocuments.us/reader035/viewer/2022062321/56649e2b5503460f94b19062/html5/thumbnails/152.jpg)
The ASR is used in the ISS Database
ASR / ISS