Object Categorization: The Constellation Model by Li Fei-Fei and Collaborators - Prof. Xiu, Study notes of Electrical and Electronics Engineering

Download Object Categorization: The Constellation Model by Li Fei-Fei and Collaborators - Prof. Xiu and more Study notes Electrical and Electronics Engineering in PDF only on Docsity!

Object categorization: the constellation models

Li Fei-Fei

with many thanks to Rob Fergus^ with many thanks to Rob Fergus

The People and slides creditPietro Perona Mike Burl^

Thomas Leung

Markus WeberRob Fergus

Andrew Zisserman^ Max Welling Li Fei-Fei

• Challenges 1: view point variation Michelangelo 1475-

Challenges 2: illumination

slide credit: S. Ullman

Challenges 4: scale

Challenges 5: deformation

Xu, Beihong 1943

Challenges 7: intra-class variation

Goal

-^ Recognition of visual object classes •^ Unassisted learning

Model: Parts and Structure

Parts and Structure Literature• Fischler & Elschlager 1973 • Yuille ‘91 • Brunelli & Poggio ‘93 • Lades, v.d. Malsburg et al. ‘93 • Cootes, Lanitis, Taylor et al. ‘95 • Amit & Geman ‘95, ‘99 • et al. Perona ‘95, ‘96, ’98, ’00, ‘03 • Huttenlocher et al. ’00 • Agarwal & Roth ’02 etc…

A^

B^ D

Deformations C

Presence / Absence of Features

occlusion

Foreground model

Generative probabilistic model Gaussian shape pdf

Clutter model^ Uniform shape pdf Prob. of detection^ 0.^

0.75 0.

# detections p ( N | Poisson^2

λ)^2

p ( N | Poisson^1

λ)^1

p ( N | Poisson^3

λ)^3

Assumptions: (a) Clutter independent of foreground detections

(b) Clutter detections independent of each other

Example 1. Object Part Positions

3a. N false detect

2. Part Absence

N^1 N^2

3b. Position f. detect N^3

Learning Models `Manually’ • Obtain set of training images • Choose parts • Label parts by hand, train detectors• Learn model from labeled parts