4.

Recognition
COMPUTER VISION – TK44019
 Definition
• Recognition = to know again
• Learning:
– Supervised(Classification)
• Classes: known by some description/set of examples
• Sensor/transducer: sense the actual physical and output
a digital representation
• Feature extractor: extracts the relevant information
• Classifier: assigns the object (extracted feature) to one of
the designated classes
– Unsupervised (Clustering)
 What is Clustering?
• Organizing data into classes such that there is
• high intra-class similarity

• low inter-class similarity

• Finding the class labels and the number of classes directly

from the data (in contrast to classification).
• More informally, finding natural groupings among objects.
What is a natural grouping among these objects?
What is a natural grouping among these objects?

Clustering is subjective

Simpson's Family School Employees Females Males

 What is Similarity?
The quality or state of being similar; likeness; resemblance; as, a similarity of features.
Webster's Dictionary

Similarity is hard
to define, but…
“We know it when
we see it”

The real meaning

of similarity is a
philosophical
question. We will
take a more
pragmatic
approach.
 Defining Distance Measures
Definition: Let O1 and O2 be two objects from the
universe of possible objects. The distance (dissimilarity)
between O1 and O2 is a real number denoted by D(O1,O2)

Peter Piotr

0.23 3 342.7
Peter Piotr When we peek inside one of
these black boxes, we see some
function on two variables. These
functions might very simple or
very complex.
In either case it is natural to ask,
what properties should these
functions have?
3

What properties should a distance measure have?

• D(A,B) = D(B,A) Symmetry

• D(A,A) = 0 Constancy of Self-Similarity
• D(A,B) = 0 IIf A= B Positivity (Separation)
• D(A,B) ≤ D(A,C) + D(B,C) Triangular Inequality
 Intuitions behind desirable
distance measure properties
D(A,B) = D(B,A) Symmetry
Otherwise you could claim “Alex looks like Bob, but Bob looks nothing like Alex.”

D(A,A) = 0 Constancy of Self-Similarity

Otherwise you could claim “Alex looks more like Bob, than Bob does.”

D(A,B) = 0 IIf A=B Positivity (Separation)

Otherwise there are objects in your world that are different, but you cannot tell apart.

D(A,B) ≤ D(A,C) + D(B,C) Triangular Inequality

Otherwise you could claim “Alex is very like Bob, and Alex is very like Carl, but Bob
is very unlike Carl.”
 Two Types of Clustering
• Partitional algorithms: Construct various partitions and then
evaluate them by some criterion (we will see an example called BIRCH)
• Hierarchical algorithms: Create a hierarchical decomposition of
the set of objects using some criterion

Hierarchical Partitional
Desirable Properties of a Clustering Algorithm

• Scalability (in terms of both time and space)

• Ability to deal with different data types
• Minimal requirements for domain knowledge to
determine input parameters
• Able to deal with noise and outliers
• Insensitive to order of input records
• Incorporation of user-specified constraints
• Interpretability and usability
A Useful Tool for Summarizing Similarity Measurements
In order to better appreciate and evaluate the examples given in the
early part of this talk, we will now introduce the dendrogram.

Terminal Branch Root

The similarity between two objects in a
Internal Branch
Internal Node
dendrogram is represented as the height of
Leaf the lowest internal node they share.
 A Demonstration of Hierarchical Clustering using String Edit Distance
Pedro (Portuguese)
Petros (Greek), Peter (English), Piotr (Polish), Peadar
(Irish), Pierre (French), Peder (Danish), Peka
(Hawaiian), Pietro (Italian), Piero (Italian Alternative),
Petr (Czech), Pyotr (Russian)

Cristovao (Portuguese)
Christoph (German), Christophe (French), Cristobal
(Spanish), Cristoforo (Italian), Kristoffer
(Scandinavian), Krystof (Czech), Christopher (English)

Miguel (Portuguese)
Michalis (Greek), Michael (English), Mick (Irish!)
 Hierarchal clustering can sometimes show
patterns that are meaningless or spurious
• For example, in this clustering, the tight grouping of Australia,
Anguilla, St. Helena etc is meaningful, since all these countries are
former UK colonies.

• However the tight grouping of Niger and India is completely

spurious, there is no connection between the two.

South Georgia & Serbia &

St. Helena & South Sandwich Montenegro
AUSTRALIA Dependencies ANGUILLA Islands U.K. (Yugoslavia) FRANCE NIGER INDIA IRELAND BRAZIL
We can look at the dendrogram to determine the “correct” number of
clusters. In this case, the two highly separated subtrees are highly
suggestive of two clusters. (Things are rarely this clear cut, unfortunately)
One potential use of a dendrogram is to detect outliers
The single isolated branch is suggestive of a
data point that is very different to all others

Outlier
 (How-to) Hierarchical Clustering
The number of dendrograms with n Since we cannot test all possible trees
leafs = (2n -3)!/[(2(n -2)) (n -2)!] we will have to heuristic search of all
possible trees. We could do this..
Number Number of Possible
of Leafs Dendrograms
2 1 Bottom-Up (agglomerative): Starting
3 3
4 15
with each item in its own cluster, find
5 105 the best pair to merge into a new
... …
10 34,459,425
cluster. Repeat until all clusters are
fused together.

Top-Down (divisive): Starting with all

the data in a single cluster, consider
every possible way to divide the cluster
into two. Choose the best division and
recursively operate on both sides.
We begin with a distance
matrix which contains the
distances between every pair
of objects in our database.

0 8 8 7 7

0 2 4 4

0 3 3
D( , ) = 8 0 1

D( , ) = 1 0
 Bottom--Up ((agglomerative
Bottom agglomerative):
):
Starting with each item in its own
cluster, find the best pair to merge into
a new cluster. Repeat until all clusters
are fused together.

Consider all Choose

possible … the best
merges…
 Bottom--Up ((agglomerative
Bottom agglomerative):
):
Starting with each item in its own
cluster, find the best pair to merge into
a new cluster. Repeat until all clusters
are fused together.

Consider all
Choose
possible
… the best
merges…

Consider all Choose

possible … the best
merges…
 Bottom--Up ((agglomerative
Bottom agglomerative):
):
Starting with each item in its own
cluster, find the best pair to merge into
a new cluster. Repeat until all clusters
are fused together.

Consider all
Choose
possible
… the best
merges…

Consider all
Choose
possible
… the best
merges…

Consider all Choose

possible … the best
merges…
 Bottom--Up ((agglomerative
Bottom agglomerative):
):
Starting with each item in its own
cluster, find the best pair to merge into
a new cluster. Repeat until all clusters
are fused together.

Consider all
Choose
possible
… the best
merges…

Consider all
Choose
possible
… the best
merges…

Consider all Choose

possible … the best
merges…
 We know how to measure the distance between two
objects, but defining the distance between an object
and a cluster, or defining the distance between two
clusters is non obvious.

• Single linkage (nearest neighbor): In this method the distance

between two clusters is determined by the distance of the two closest objects
(nearest neighbors) in the different clusters.
• Complete linkage (furthest neighbor): In this method, the
distances between clusters are determined by the greatest distance between any
two objects in the different clusters (i.e., by the "furthest neighbors").
• Group average linkage: linkage In this method, the distance between two
clusters is calculated as the average distance between all pairs of objects in the
two different clusters.
• Wards Linkage:
Linkage In this method, we try to minimize the variance of the
merged clusters
 Single linkage

7
25

20

15
4

3 10

2
5

29 2 6 11 9 17 10 13 24 25 26 20 22 30 27 1 3 8 4 12 5 14 23 15 16 18 19 21 28 7 0
5 14 23 7 4 12 19 21 24 15 16 18 1 3 8 9 29 2 10 11 20 28 17 26 27 25 6 13 22 30

Average linkage Wards linkage

Summary of Hierarchal Clustering Methods

• No need to specify the number of clusters in

advance.
• Hierarchal nature maps nicely onto human intuition
for some domains
• They do not scale well: time complexity of at least
O(n2), where n is the number of total objects.
• Like any heuristic search algorithms, local optima
are a problem.
• Interpretation of results is (very) subjective.
Up to this point we have simply assumed that we can measure
similarity, but

How do we measure similarity?

Peter Piotr

0.23 3 342.7
 A generic technique for measuring similarity
To measure the similarity between two objects,
transform one of the objects into the other, and
measure how much effort it took. The measure
of effort becomes the distance measure.

The distance between Patty and Selma.

Change dress color, 1 point
Change earring shape, 1 point
Change hair part, 1 point
D(Patty,Selma) = 3

The distance between Marge and Selma.

Change dress color, 1 point
Add earrings, 1 point This is called the “edit
Decrease height, 1 point distance” or the
Take up smoking, 1 point “transformation distance”
Lose weight, 1 point
D(Marge,Selma) = 5
Edit Distance Example How similar are the names
It is possible to transform any string Q into “Peter” and “Piotr”?
Assume the following cost function
string C, using only Substitution, Insertion Substitution 1 Unit
and Deletion. Insertion 1 Unit
Assume that each of these operators has a Deletion 1 Unit
cost associated with it.
D(Peter,Piotr) is 3
The similarity between two strings can be
defined as the cost of the cheapest
transformation from Q to C. Peter
Note that for now we have ignored the issue of how we can find this cheapest
transformation Substitution (i for e)
Piter
Insertion (o)

Pioter
Deletion (e)

Piotr
 Partitional Clustering
• Nonhierarchical, each instance is placed in
exactly one of K nonoverlapping clusters.
• The user normally has to input the desired
number of clusters K.
Algorithm k-means
1. Decide on a value for k.
2. Initialize the k cluster centers (randomly, if
necessary).
3. Decide the class memberships of the N objects by
assigning them to the nearest cluster center.
4. Re-estimate the k cluster centers, by assuming the
memberships found above are correct.
5. If none of the N objects changed membership in
the last iteration, exit. Otherwise goto 3.
 K-means Clustering: Step 1
Algorithm: k-means, Distance Metric: Euclidean Distance
5

4
k1
3

2
k2

k3
0
0 1 2 3 4 5
 K-means Clustering: Step 2
Algorithm: k-means, Distance Metric: Euclidean Distance
5

4
k1
3

2
k2

k3
0
0 1 2 3 4 5
 K-means Clustering: Step 3
Algorithm: k-means, Distance Metric: Euclidean Distance
5

4
k1

k3
1 k2

0
0 1 2 3 4 5
 K-means Clustering: Step 4
Algorithm: k-means, Distance Metric: Euclidean Distance
5

4
k1

k3
1 k2

0
0 1 2 3 4 5
 K-means Clustering: Step 5
Algorithm: k-means, Distance Metric: Euclidean Distance
5

expression in condition 2
4
k1
3

k2
1 k3

0
0 1 2 3 4 5

expression in condition 1
 Comments on the K-Means Method
• Strength
– Relatively efficient: O(tkn), where n is # objects, k is # clusters,
and t is # iterations. Normally, k, t << n.
– Often terminates at a local optimum. The global optimum may
be found using techniques such as: deterministic annealing and
genetic algorithms
• Weakness
– Applicable only when mean is defined, then what about
categorical data?
– Need to specify k, the number of clusters, in advance
– Unable to handle noisy data and outliers
– Not suitable to discover clusters with non-convex shapes
 The K-Medoids Clustering Method
• Find representative objects, called medoids, in clusters
• PAM (Partitioning Around Medoids, 1987)
– starts from an initial set of medoids and iteratively replaces
one of the medoids by one of the non-medoids if it improves
the total distance of the resulting clustering
– PAM works effectively for small data sets, but does not scale
well for large data sets
 What happens if the data is streaming…

Nearest Neighbor Clustering

Not to be confused with Nearest Neighbor Classification

• Items are iteratively merged into the

existing clusters that are closest.
• Incremental
• Threshold, t, used to determine if items are
added to existing clusters or a new cluster is
created.
 10
9
8
7
6
Threshold t 5
4
3
1
t 2
1
2
1 2 3 4 5 6 7 8 9 10
 10
9
8
7
6

New data point arrives… 5

4

It is within the threshold for 3

1
2 3
cluster 1, so add it to the
cluster, and update cluster 1
2
center.
1 2 3 4 5 6 7 8 9 10
 10
New data point arrives…
9 4
It is not within the threshold 8

for cluster 1, so create a new 7

cluster, and so on.. 6
5
4
3
1
2 3
1
2

Algorithm is highly order 1 2 3 4 5 6 7 8 9 10

dependent…

It is difficult to determine t in
advance…
 How can we tell the right number of clusters?

In general, this is a unsolved problem. However there are many

approximate methods, such as “knee finding” or “elbow finding”.
10
9
8
7
6
5
4
3
2
1

1 2 3 4 5 6 7 8 9 10