Sliding windows and face detection Tuesday, Nov 10 Kristen Grauman UT - - PDF document

▶
sliding windows and face detection
SMART_READER_LITE
LIVE PREVIEW

Sliding windows and face detection Tuesday, Nov 10 Kristen Grauman UT - - PDF document

Mar 06, 2023 526 likes •943 views

11/10/2009 Sliding windows and face detection Tuesday, Nov 10 Kristen Grauman UT Austin Last time Modeling categories with local features and spatial information: Histograms, configurations of visual words to capture Histograms

slide-1
SLIDE 1

11/10/2009 1

Sliding windows and face detection

Tuesday, Nov 10 Kristen Grauman UT‐Austin

Last time

  • Modeling categories with local features and

spatial information:

Histograms configurations of visual words to capture – Histograms, configurations of visual words to capture global or local layout in the bag-of-words framework

  • Pyramid match, semi-local features
slide-2
SLIDE 2

11/10/2009 2

Pyramid match

Histogram intersection counts number of possible matches at a given partitioning.

  • Make a pyramid of bag‐of‐words histograms.
  • Provides some loose (global) spatial layout information

Spatial pyramid match

[Lazebnik, S chmid & Ponce, CVPR 2006]

slide-3
SLIDE 3

11/10/2009 3

Last time

  • Modeling categories with local features and

spatial information:

Histograms configurations of visual words to capture – Histograms, configurations of visual words to capture global or local layout in the bag-of-words framework

  • Pyramid match, semi-local features

– Part-based models to encode category’s part appearance together with 2d layout, – Allow detection within cluttered image Allow detection within cluttered image

  • “implicit shape model”, Generalized Hough for detection
  • “constellation model”: exhaustive search for best fit of features

to parts

Implicit shape models

  • Visual vocabulary is used to index votes for
  • bject position [a visual word = “part”]
  • B. Leibe, A. Leonardis, and B. Schiele, Combined Object Categorization and

Segmentation with an Implicit Shape Model, ECCV Workshop on Statistical Learning in Computer Vision 2004 visual codeword with displacement vectors training image annotated with object localization info

slide-4
SLIDE 4

11/10/2009 4

Implicit shape models

  • Visual vocabulary is used to index votes for
  • bject position [a visual word = “part”]
  • B. Leibe, A. Leonardis, and B. Schiele, Combined Object Categorization and

Segmentation with an Implicit Shape Model, ECCV Workshop on Statistical Learning in Computer Vision 2004 test image

ng

Shape representation in part-based models

Fully connected constellation model “Star” shape model

  • ry Augmented Computi

gnition Tutorial

x1 x3 x4 x6 x5 x2

e g Constellation Model

x1 x3 x4 x6 x5 x2 i li it h d l

Perceptual and Sens

Visual Object Recog

e.g. Constellation Model Parts fully connected Recognition complexity: O(NP) Method: Exhaustive search e.g. implicit shape model Parts mutually independent Recognition complexity: O(NP) Method: Gen. Hough Transform

S lide credit: Rob Fergus

N image features, P parts in the model

slide-5
SLIDE 5

11/10/2009 5

Coarse genres of recognition approaches

  • Alignment: hypothesize and test

Pose clustering with object instances – Pose clustering with object instances – Indexing invariant features + verification

  • Local features: as parts or words

– Part-based models – Bags of words models g

  • Global appearance: “texture templates”

– With or without a sliding window

Today

  • Detection as classification

– Supervised classification

  • Skin color detection example

– Sliding window detection

  • Face detection example
slide-6
SLIDE 6

11/10/2009 6

Supervised classification

  • Given a collection of labeled examples, come up with a

function that will predict the labels of new examples.

“four” “nine”

?

Training examples Novel input

  • How good is some function we come up with to do the

classification?

  • Depends on

– Mistakes made – Cost associated with the mistakes

Supervised classification

  • Given a collection of labeled examples, come up with a

function that will predict the labels of new examples.

  • Consider the two-class (binary) decision problem

– L(4→9): Loss of classifying a 4 as a 9 – L(9→4): Loss of classifying a 9 as a 4

  • Risk of a classifier s is expected loss:
  • We want to choose a classifier so as to minimize this

total risk

( ) ( ) ( ) ( )

4 9 using | 4 9 Pr 9 4 using | 9 4 Pr ) ( → → + → → = L s L s s R

slide-7
SLIDE 7

11/10/2009 7

Supervised classification

Optimal classifier will minimize total risk.

Feature value x

At decision boundary, either choice of label yields same expected loss. If we choose class “four” at boundary, expected loss is: If we choose class “nine” at boundary, expected loss is:

4) (9 ) | 9 is class ( 4) (4 ) | 4 is (class 4) (9 ) | 9 is class ( → = → + → = L P L P L P x x x 9) (4 ) | 4 is class ( → = L P x

Supervised classification

Optimal classifier will minimize total risk.

Feature value x

At decision boundary, either choice of label yields same expected loss. So, best decision boundary is at point x where To classify a new point, choose class with lowest expected loss; i.e., choose “four” if

9) (4 ) | 4 is P(class 4) (9 ) | 9 is class ( → = → L L P x x

) 4 9 ( ) | 9 ( ) 9 4 ( ) | 4 ( → > → L P L P x x

slide-8
SLIDE 8

11/10/2009 8

Supervised classification

Optimal classifier will minimize total risk.

P(4 | x) P(9 | x)

Feature value x

At decision boundary, either choice of label yields same expected loss. So, best decision boundary is at point x where To classify a new point, choose class with lowest expected loss; i.e., choose “four” if

9) (4 ) | 4 is P(class 4) (9 ) | 9 is class ( → = → L L P x x

) 4 9 ( ) | 9 ( ) 9 4 ( ) | 4 ( → > → L P L P x x

How to evaluate these probabilities?

Probability

Basic probability

  • X is a random variable
  • P(X) is the probability that X achieves a certain value

ll d PDF

  • called a PDF
  • probability distribution/density function
  • r
  • Conditional probability: P(X | Y)

– probability of X given that we already know Y continuous X discrete X

Source: Steve Seitz

slide-9
SLIDE 9

11/10/2009 9

Example: learning skin colors

  • We can represent a class-conditional density using a

histogram (a “non-parametric” distribution)

Feature x = Hue P(x|skin)

Percentage of skin pixels in each bin

Feature x = Hue P(x|not skin)

Example: learning skin colors

  • We can represent a class-conditional density using a

histogram (a “non-parametric” distribution)

Feature x = Hue P(x|skin) N t i Feature x = Hue P(x|not skin) Now we get a new image, and want to label each pixel as skin or non-skin. What’s the probability we care about to do skin detection?

slide-10
SLIDE 10

11/10/2009 10

Bayes rule

) ( ) | ( skin P skin x P

posterior prior likelihood

) ( ) ( ) | ( ) | ( x P skin P skin x P x skin P = ) ( ) | ( ) | ( skin P skin x P x skin P α ) ( ) | ( ) | (

Where does the prior come from? Why use a prior?

Example: classifying skin pixels

Now for every pixel in a new image, we can estimate probability that it is generated by skin.

Brighter pixels higher probability

  • f being skin

Classify pixels based on these probabilities

slide-11
SLIDE 11

11/10/2009 11

Example: classifying skin pixels

Gary Bradski, 1998

Example: classifying skin pixels

Gary Bradski, 1998

Using skin color-based face detection and pose estimation as a video-based interface

slide-12
SLIDE 12

11/10/2009 12

Supervised classification

  • Want to minimize the expected misclassification

T l t t i

  • Two general strategies

– Use the training data to build representative probability model; separately model class-conditional densities and priors (generative) – Directly construct a good decision boundary, model the posterior (discriminative)

Today

  • Detection as classification

– Supervised classification

  • Skin color detection example

– Sliding window detection

  • Face detection example
slide-13
SLIDE 13

11/10/2009 13

ng

Detection via classification: Main idea

Basic component: a binary classifier

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Car/non-car Classifier

Perceptual and Sens

Visual Object Recog Visual Object Recog

Yes, car. No, not a car.

ng

Detection via classification: Main idea

If object may be in a cluttered scene, slide a window around looking for it.

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Car/non-car Classifier

Perceptual and Sens

Visual Object Recog Visual Object Recog

(Essentially, our skin detector was doing this, with a window that was one pixel big.)

slide-14
SLIDE 14

11/10/2009 14

ng

Detection via classification: Main idea

Fleshing out this pipeline a bit more, we need to:

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Training examples 1. Obtain training data 2. Define features 3. Define classifier

Perceptual and Sens

Visual Object Recog Visual Object Recog

Car/non-car Classifier Feature extraction

ng

Detection via classification: Main idea

  • Consider all subwindows in an image

Sample at multiple scales and positions (and orientations)

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

  • Make a decision per window:

“Does this contain object category X or not?”

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-15
SLIDE 15

11/10/2009 15

ng

Feature extraction: global appearance

Feature extraction

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Simple holistic descriptions of image content

Perceptual and Sens

Visual Object Recog Visual Object Recog

grayscale / color histogram vector of pixel intensities

ng

Feature extraction: global appearance

  • Pixel-based representations sensitive to small shifts
  • ry Augmented Computi

gnition Tutorial gnition Tutorial

  • Color or grayscale-based appearance description can be

sensitive to illumination and intra-class appearance variation

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-16
SLIDE 16

11/10/2009 16

ng

Gradient-based representations

  • Consider edges, contours, and (oriented) intensity

gradients

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

ng

Gradient-based representations

  • Consider edges, contours, and (oriented) intensity

gradients

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

  • Summarize local distribution of gradients with histogram

Locally orderless: offers invariance to small shifts and rotations Contrast-normalization: try to correct for variable illumination

slide-17
SLIDE 17

11/10/2009 17

ng

  • How to compute a decision for each

subwindow?

Classifier construction

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

subwindow?

Image feature

Perceptual and Sens

Visual Object Recog Visual Object Recog

g

ng

Discriminative classifier construction: many choices…

Nearest neighbor Neural networks

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

106 examples

Shakhnarovich, Viola, Darrell 2003 Berg, Berg, Malik 2005... LeCun, Bottou, Bengio, Haffner 1998 Rowley, Baluja, Kanade 1998 … Support Vector Machines Conditional Random Fields Boosting

Perceptual and Sens

Visual Object Recog Visual Object Recog

McCallum, Freitag, Pereira 2000; Kumar, Hebert 2003 … Guyon, Vapnik Heisele, Serre, Poggio, 2001,…

S lide adapted from Antonio Torralba

  • K. Grauman, B. Leibe

Viola, Jones 2001, Torralba et al. 2004, Opelt et al. 2006,…

slide-18
SLIDE 18

11/10/2009 18

ng

Boosting

  • Build a strong classifier by combining number of “weak

classifiers”, which need only be better than chance

  • Sequential learning process: at each iteration add a
  • ry Augmented Computi

gnition Tutorial gnition Tutorial

  • Sequential learning process: at each iteration, add a

weak classifier

  • Flexible to choice of weak learner

including fast simple classifiers that alone may be inaccurate

  • We’ll look at the AdaBoost algorithm

Perceptual and Sens

Visual Object Recog Visual Object Recog

Easy to implement Base learning algorithm for Viola-Jones face detector

ng

AdaBoost: Intuition

Consider a 2-d feature space with positive and i l

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

negative examples. Each weak classifier splits the training examples with at least 50% accuracy. Examples misclassified by i k l

Perceptual and Sens

Visual Object Recog Visual Object Recog

Figure adapted from Freund and S chapire

a previous weak learner are given more emphasis at future rounds.

slide-19
SLIDE 19

11/10/2009 19

ng

AdaBoost: Intuition

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

ng

AdaBoost: Intuition

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-20
SLIDE 20

11/10/2009 20

ng

AdaBoost: Intuition

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

Final classifier is combination of the weak classifiers

ng

Boosting: Training procedure

  • Initially, weight each training example equally
  • In each boosting round:
  • Find the weak learner that achieves the lowest weighted
  • ry Augmented Computi

gnition Tutorial gnition Tutorial

training error

  • Raise the weights of training examples misclassified by

current weak learner

  • Compute final classifier as linear combination of all

weak learners (weight of each learner is directly proportional to its accuracy) E t f l f i hti d bi i

Perceptual and Sens

Visual Object Recog Visual Object Recog

  • Exact formulas for re-weighting and combining

weak learners depend on the particular boosting scheme (e.g., AdaBoost)

S lide credit: Lana Lazebnik

slide-21
SLIDE 21

11/10/2009 21

ng

AdaBoost Algorithm

S tart with uniform weights

  • n training

examples

{x1,… xn}

d

  • ry Augmented Computi

gnition Tutorial gnition Tutorial Evaluate

weight ed error for each feature, pick best. Re-weight the examples: For T rounds

Perceptual and Sens

Visual Object Recog Visual Object Recog Incorrectly classified -> more weight Correctly classified -> less weight Final classifier is combination of the weak ones, weighted according to error they had. Freund & Schapire 1995

ng

Faces : terminology

  • Detection: given an
  • ry Augmented Computi

gnition Tutorial gnition Tutorial

  • Detection: given an

image, where is the face? Recognition: hose

Perceptual and Sens

Visual Object Recog Visual Object Recog

  • Recognition: whose

face is it?

Image credit: H. Rowley

Ann

slide-22
SLIDE 22

11/10/2009 22

ng

Example: Face detection

  • Frontal faces are a good example of a class where

global appearance models + a sliding window detection approach fit well:

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

detection approach fit well:

Regular 2D structure Center of face almost shaped like a “patch”/window

Perceptual and Sens

Visual Object Recog Visual Object Recog

  • Now we’ll take AdaBoost and see how the Viola-

Jones face detector works

ng

Feature extraction

Feature output is difference between adjacent regions “Rectangular” filters

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Efficiently computable with integral image: any sum can be computed

Value at (x,y) is sum of pixels above and to the left of (x,y) Perceptual and Sens

Visual Object Recog Visual Object Recog Viola & Jones, CVPR 2001

in constant time Avoid scaling images scale features directly for same cost

Integral image

slide-23
SLIDE 23

11/10/2009 23

ng

Large library of filters

Considering all possible filter parameters:

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

p position, scale, and type: 180,000+ possible features associated with each 24 x 24 window

Perceptual and Sens

Visual Object Recog Visual Object Recog

window Which subset of these features should we use to determine if a window has a face? Use AdaBoost both to select the informative features and to form the classifier

ng

AdaBoost for feature+classifier selection

  • Want to select the single rectangle feature and threshold

that best separates positive (faces) and negative (non- faces) training examples, in terms of weighted error.

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Resulting weak classifier:

Perceptual and Sens

Visual Object Recog Visual Object Recog Outputs of a possible rectangle feature on faces and non-faces.

… For next round, reweight the examples according to errors, choose another filter/threshold combo.

slide-24
SLIDE 24

11/10/2009 24

  • Even if the filters are fast to compute, each

i h l t f ibl i d t new image has a lot of possible windows to search.

  • How to make the detection more efficient?

ng

Cascading classifiers for detection

For efficiency, apply less accurate but faster classifiers first to immediately discard

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

first to immediately discard windows that clearly appear to be negative; e.g.,

  • Filter for promising regions with an

initial inexpensive classifier

  • Build a chain of classifiers, choosing

cheap ones with low false negative

Perceptual and Sens

Visual Object Recog Visual Object Recog

p g rates early in the chain

Figure from Viola & Jones CVPR 2001

slide-25
SLIDE 25

11/10/2009 25

ng

Viola-Jones Face Detector: Summary

Train cascade of classifiers with Ad B t

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Faces Non-faces

AdaBoost

Selected features, thresholds, and weights New image

Perceptual and Sens

Visual Object Recog Visual Object Recog

  • Train with 5K positives, 350M negatives
  • Real-time detector using 38 layer cascade
  • 6061 features in final layer
  • [Implementation available in OpenCV:

http://www.intel.com/technology/computing/opencv/]

ng

Viola-Jones Face Detector: Summary

  • A seminal approach to real-time object detection
  • Training is slow, but detection is very fast

Key ideas

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

  • Key ideas

Integral images for fast feature evaluation Boosting for feature selection Attentional cascade for fast rejection of non-face windows

Perceptual and Sens

Visual Object Recog Visual Object Recog

P . Viola and M. Jones. Rapid obj ect det ect ion using a boost ed cascade of simple feat ures. CVPR 2001. P . Viola and M. Jones. Robust real-t ime face det ect ion. IJCV 57(2), 2004.

S lide credit: Lana Lazebnik

slide-26
SLIDE 26

11/10/2009 26

ng

First two features

Viola-Jones Face Detector: Results

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

First two features selected

Perceptual and Sens

Visual Object Recog Visual Object Recog

ng

Viola-Jones Face Detector: Results

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-27
SLIDE 27

11/10/2009 27

ng

Viola-Jones Face Detector: Results

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

ng

Viola-Jones Face Detector: Results

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-28
SLIDE 28

11/10/2009 28

ng

Detecting profile faces?

Can we use the same detector?

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

ng

Viola-Jones Face Detector: Results

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

Paul Viola, ICCV tutorial

slide-29
SLIDE 29

11/10/2009 29

ng

Example application

Frontal faces detected and

  • ry Augmented Computi

gnition Tutorial gnition Tutorial detected and then tracked, character names inferred with alignment of script and subtitles.

Perceptual and Sens

Visual Object Recog Visual Object Recog Everingham, M., Sivic, J. and Zisserman, A. "Hello! My name is... Buffy" - Automatic naming of characters in TV video, BMVC 2006. http:/ / www.robots.ox.ac.uk/ ~vgg/ research/ nface/ index.html

ng

Example application: faces in photos

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-30
SLIDE 30

11/10/2009 30

Consumer application: iPhoto 2009

http://www.apple.com/ilife/iphoto/

Slide credit: Lana Lazebnik

Consumer application: iPhoto 2009

Can be trained to recognize pets!

http://www.maclife.com/article/news/iphotos_faces_recognizes_cats

Slide credit: Lana Lazebnik

slide-31
SLIDE 31

11/10/2009 31

Consumer application: iPhoto 2009

Things iPhoto thinks are faces

Slide credit: Lana Lazebnik

  • Other classes that might work with global

i i d ? appearance in a window?

slide-32
SLIDE 32

11/10/2009 32

ng

Pedestrian detection

  • Detecting upright, walking humans also possible using sliding

window’s appearance/texture; e.g.,

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

SVM with Haar wavelets [Papageorgiou & Poggio, IJCV 2000] Space-time rectangle features [Viola, Jones & SVM with HoGs [Dalal & Triggs, CVPR 2005]

Perceptual and Sens

Visual Object Recog Visual Object Recog

2000] [ , Snow, ICCV 2003] gg , ]

  • Other classes that might work with global

i i d ? appearance in a window?

slide-33
SLIDE 33

11/10/2009 33

Penguin detection & identification

Burghart, Thomas, Barham, and Calic. Automated Visual Recognition of Individual African Penguins , 2004.

This project uses the Viola‐Jones Adaboost face detection algorithm to detect penguin chests, and then matches the pattern of spots to identify a particular penguin.

slide-34
SLIDE 34

11/10/2009 34

Use rectangular features, select good features to distinguish the chest from non‐chests with Adaboost

Burghart, Thomas, Barham, and Calic. Automated Visual Recognition of Individual African Penguins , 2004.

Attentional cascade Penguin chest detections

Burghart, Thomas, Barham, and Calic. Automated Visual Recognition of Individual African Penguins , 2004.

slide-35
SLIDE 35

11/10/2009 35

Given a detected chest, try to extract the whole chest for this particular penguin.

Burghart, Thomas, Barham, and Calic. Automated Visual Recognition of Individual African Penguins , 2004.

Example detections

Burghart, Thomas, Barham, and Calic. Automated Visual Recognition of Individual African Penguins , 2004.

slide-36
SLIDE 36

11/10/2009 36

Perform identification by matching the pattern of spots to a database of known penguins.

Burghart, Thomas, Barham, and Calic. Automated Visual Recognition of Individual African Penguins , 2004.

Penguin detection & identification

Burghart, Thomas, Barham, and Calic. Automated Visual Recognition of Individual African Penguins , 2004.

slide-37
SLIDE 37

11/10/2009 37

ng

Highlights

  • Sliding window detection and global appearance

descriptors:

Simple detection protocol to implement

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Simple detection protocol to implement Good feature choices critical Past successes for certain classes

Perceptual and Sens

Visual Object Recog Visual Object Recog

ng

Limitations

  • High computational complexity

For example: 250,000 locations x 30 orientations x 4 scales =

30,000,000 evaluations!

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

If training binary detectors independently, means cost increases

linearly with number of classes

  • With so many windows, false positive rate better be low

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-38
SLIDE 38

11/10/2009 38

ng

Limitations (continued)

  • Not all objects are “box” shaped
  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Perceptual and Sens

Visual Object Recog Visual Object Recog

ng

Limitations (continued)

  • Non-rigid, deformable objects not captured well with

representations assuming a fixed 2d structure; or must assume fixed viewpoint

  • ry Augmented Computi

gnition Tutorial gnition Tutorial

assume fixed viewpoint

  • Objects with less-regular textures not captured well

with holistic appearance-based descriptions

Perceptual and Sens

Visual Object Recog Visual Object Recog

slide-39
SLIDE 39

11/10/2009 39

ng

Limitations (continued)

  • If considering windows in isolation, context is lost
  • ry Augmented Computi

gnition Tutorial gnition Tutorial

Sliding window Detector’s view

Perceptual and Sens

Visual Object Recog Visual Object Recog

Figure credit: Derek Hoiem

ng

Limitations (continued)

  • In practice, often entails large, cropped training set

(expensive)

  • Requiring good match to a global appearance description
  • ry Augmented Computi

gnition Tutorial gnition Tutorial

  • Requiring good match to a global appearance description

can lead to sensitivity to partial occlusions

Perceptual and Sens

Visual Object Recog Visual Object Recog

Image credit: Adam, Rivlin, & S himshoni

slide-40
SLIDE 40

11/10/2009 40

Summary: Detection as classification

– Supervised classification

  • Loss and risk, Bayes rule
  • Skin color detection example

– Sliding window detection

  • Classifiers, boosting algorithm, cascades
  • Face detection example

– Limitations of a global appearance description – Limitations of sliding window detectors