Neuronale Netze

1
Neuronale Netze

• III. Mustererkennung

2
Übersicht

• Einführung
• Das visuelle System
• David Marr
• The three levels
• Representational framework
• Repräsentation von Mustern
• Muster als Vektoren
• Mustererkennung
• Klassifikationstechniken

3
1. Einführung
4
2. Das visuelle System
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
Stäbchen sehen nur Grautöne,Drei Zapfentypen
teilen sich das sichtbare Spektrum
Wellenlänge des Lichts in nm
9
3. David Marr (1945-1980)

• ? 1982 Vision A Computational Investigation
  into the Human Representation and Processing of
  Visual Information

10
3.1 The three levels
11
Beispiel Registrierkasse

• Constraints
• If you buy nothing, it should cost you nothing
  and buying nothing and something should cost the
  same as buying just the something (The rules for
  zero).
• The order in which the goods are presented to the
  cashier should not affect the total
  (Commutativity).
• Arranging the goods into two piles and paying for
  each pile separately should not affect the total
  amount you pay (Associativity)
• If you buy an item and then return it for a
  refund, your total expenditure should be zero
  (Inverses).

12
3.2 Representational framework
13
(No Transcript)
14
(No Transcript)
15
4. Repräsentation von Mustern
16
4.1 Muster als Vektoren
17
4.2 Mustererkennung

• (pattern recognition) is a search for structure
  in data Bezdeck, 1981
• Daten
• Musterräume
• Merkmale
• Klassifikatoren
• Entscheidungsgrenzen

18
4.3 Klassifikationstechniken
19
4.3.1 nearest neighbour

• In pattern recognition, the k-nearest neighbor
  algorithm (k-NN) is a method for classifying
  objects based on closest training examples in the
  feature space. k-NN is a type of instance-based
  learning, or lazy-learning where the function is
  only approximated locally and all computation is
  deferred until classification.

20
4.3.2 Distanzmetriken (Hamming, Euklid)
21
4.3.3 Lineare Klassifikatoren
22
4.3.4 Bayes-Klassifikation