part 1: bag-of-words models by li fei-fei (princeton)
TRANSCRIPT
![Page 1: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/1.jpg)
Part 1: Bag-of-words models
by Li Fei-Fei (Princeton)
![Page 2: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/2.jpg)
Related worksRelated works• Early “bag of words” models: mostly texture
recognition– Cula & Dana, 2001; Leung & Malik 2001; Mori, Belongie & Malik,
2001; Schmid 2001; Varma & Zisserman, 2002, 2003; Lazebnik, Schmid & Ponce, 2003;
• Hierarchical Bayesian models for documents (pLSA, LDA, etc.)– Hoffman 1999; Blei, Ng & Jordan, 2004; Teh, Jordan, Beal &
Blei, 2004
• Object categorization– Csurka, Bray, Dance & Fan, 2004; Sivic, Russell, Efros,
Freeman & Zisserman, 2005; Sudderth, Torralba, Freeman & Willsky, 2005;
• Natural scene categorization– Vogel & Schiele, 2004; Fei-Fei & Perona, 2005; Bosch,
Zisserman & Munoz, 2006
![Page 3: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/3.jpg)
ObjectObject Bag of ‘words’Bag of ‘words’
![Page 4: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/4.jpg)
Analogy to documentsAnalogy to documents
Of all the sensory impressions proceeding to the brain, the visual experiences are the dominant ones. Our perception of the world around us is based essentially on the messages that reach the brain from our eyes. For a long time it was thought that the retinal image was transmitted point by point to visual centers in the brain; the cerebral cortex was a movie screen, so to speak, upon which the image in the eye was projected. Through the discoveries of Hubel and Wiesel we now know that behind the origin of the visual perception in the brain there is a considerably more complicated course of events. By following the visual impulses along their path to the various cell layers of the optical cortex, Hubel and Wiesel have been able to demonstrate that the message about the image falling on the retina undergoes a step-wise analysis in a system of nerve cells stored in columns. In this system each cell has its specific function and is responsible for a specific detail in the pattern of the retinal image.
sensory, brain, visual, perception,
retinal, cerebral cortex,eye, cell, optical
nerve, imageHubel, Wiesel
China is forecasting a trade surplus of $90bn (£51bn) to $100bn this year, a threefold increase on 2004's $32bn. The Commerce Ministry said the surplus would be created by a predicted 30% jump in exports to $750bn, compared with a 18% rise in imports to $660bn. The figures are likely to further annoy the US, which has long argued that China's exports are unfairly helped by a deliberately undervalued yuan. Beijing agrees the surplus is too high, but says the yuan is only one factor. Bank of China governor Zhou Xiaochuan said the country also needed to do more to boost domestic demand so more goods stayed within the country. China increased the value of the yuan against the dollar by 2.1% in July and permitted it to trade within a narrow band, but the US wants the yuan to be allowed to trade freely. However, Beijing has made it clear that it will take its time and tread carefully before allowing the yuan to rise further in value.
China, trade, surplus, commerce,
exports, imports, US, yuan, bank, domestic,
foreign, increase, trade, value
![Page 5: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/5.jpg)
• Looser definition– Independent features
A clarification: definition of “BoW”
![Page 6: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/6.jpg)
A clarification: definition of “BoW”• Looser definition
– Independent features
• Stricter definition– Independent features – histogram representation
![Page 7: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/7.jpg)
categorycategorydecisiondecision
learninglearning
feature detection& representation
codewords dictionarycodewords dictionary
image representation
category modelscategory models(and/or) classifiers(and/or) classifiers
recognitionrecognition
![Page 8: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/8.jpg)
feature detection& representation
codewords dictionarycodewords dictionary
image representation
RepresentationRepresentation
1.1.
2.2.
3.3.
![Page 9: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/9.jpg)
1.Feature detection and representation1.Feature detection and representation
![Page 10: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/10.jpg)
1.Feature detection 1.Feature detection and representationand representation
• Regular grid– Vogel & Schiele, 2003– Fei-Fei & Perona, 2005
![Page 11: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/11.jpg)
1.Feature detection 1.Feature detection and representationand representation
• Regular grid– Vogel & Schiele, 2003– Fei-Fei & Perona, 2005
• Interest point detector– Csurka, et al. 2004– Fei-Fei & Perona, 2005– Sivic, et al. 2005
![Page 12: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/12.jpg)
1.Feature detection 1.Feature detection and representationand representation
• Regular grid– Vogel & Schiele, 2003– Fei-Fei & Perona, 2005
• Interest point detector– Csurka, Bray, Dance & Fan, 2004– Fei-Fei & Perona, 2005– Sivic, Russell, Efros, Freeman & Zisserman, 2005
• Other methods– Random sampling (Vidal-Naquet & Ullman, 2002)– Segmentation based patches (Barnard, Duygulu,
Forsyth, de Freitas, Blei, Jordan, 2003)
![Page 13: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/13.jpg)
1.Feature 1.Feature detectiondetection and and representationrepresentation
Normalize patch
Detect patches[Mikojaczyk and Schmid ’02]
[Mata, Chum, Urban & Pajdla, ’02]
[Sivic & Zisserman, ’03]
Compute SIFT
descriptor
[Lowe’99]
Slide credit: Josef Sivic
![Page 14: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/14.jpg)
…
1.Feature 1.Feature detectiondetection and and representationrepresentation
![Page 15: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/15.jpg)
2. Codewords dictionary formation2. Codewords dictionary formation
…
![Page 16: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/16.jpg)
2. Codewords dictionary formation2. Codewords dictionary formation
Vector quantization
…
Slide credit: Josef Sivic
![Page 17: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/17.jpg)
2. Codewords dictionary formation2. Codewords dictionary formation
Fei-Fei et al. 2005
![Page 18: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/18.jpg)
Image patch examples of codewordsImage patch examples of codewords
Sivic et al. 2005
![Page 19: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/19.jpg)
3. Image representation3. Image representation
…..
fre
que
ncy
codewords
![Page 20: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/20.jpg)
feature detection& representation
codewords dictionarycodewords dictionary
image representation
RepresentationRepresentation
1.1.
2.2.
3.3.
![Page 21: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/21.jpg)
categorycategorydecisiondecision
codewords dictionarycodewords dictionary
category modelscategory models(and/or) classifiers(and/or) classifiers
Learning and Learning and RecognitionRecognition
![Page 22: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/22.jpg)
category modelscategory models(and/or) classifiers(and/or) classifiers
Learning and Learning and RecognitionRecognition
1. Generative method: - graphical models
2. Discriminative method: - SVM
![Page 23: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/23.jpg)
2 generative models2 generative models
1. Naïve Bayes classifier– Csurka Bray, Dance & Fan, 2004
2. Hierarchical Bayesian text models (pLSA and LDA)
– Background: Hoffman 2001, Blei, Ng & Jordan, 2004
– Object categorization: Sivic et al. 2005, Sudderth et al. 2005
– Natural scene categorization: Fei-Fei et al. 2005
![Page 24: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/24.jpg)
• wn: each patch in an image– wn = [0,0,…1,…,0,0]T
• w: a collection of all N patches in an image– w = [w1,w2,…,wN]
• dj: the jth image in an image collection
• c: category of the image
• z: theme or topic of the patch
First, some notationsFirst, some notations
![Page 25: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/25.jpg)
wN
c
Case #1: the Naïve Bayes modelCase #1: the Naïve Bayes model
)|()( cwpcp
Prior prob. of the object classes
Image likelihoodgiven the class
Csurka et al. 2004
N
nn cwpcp
1
)|()(
Object classdecision
)|( wcpc
c maxarg
![Page 26: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/26.jpg)
Csurka et al. 2004
![Page 27: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/27.jpg)
Csurka et al. 2004
![Page 28: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/28.jpg)
Hoffman, 2001
Case #2: Hierarchical Bayesian Case #2: Hierarchical Bayesian text modelstext models
wN
d z
D
wN
c z
D
Blei et al., 2001
Probabilistic Latent Semantic Analysis (pLSA)
Latent Dirichlet Allocation (LDA)
![Page 29: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/29.jpg)
wN
d z
D
Case #2: Hierarchical Bayesian Case #2: Hierarchical Bayesian text modelstext models
Probabilistic Latent Semantic Analysis (pLSA)
“face”
Sivic et al. ICCV 2005
![Page 30: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/30.jpg)
wN
c z
D
Case #2: Hierarchical Bayesian Case #2: Hierarchical Bayesian text modelstext models
Latent Dirichlet Allocation (LDA)
Fei-Fei et al. ICCV 2005
“beach”
![Page 31: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/31.jpg)
Case #2: the pLSA modelCase #2: the pLSA modelwN
d z
D
![Page 32: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/32.jpg)
Case #2: the pLSA modelCase #2: the pLSA modelwN
d z
D
Observed codeword distributions
Codeword distributionsper theme (topic)
Theme distributionsper image
Slide credit: Josef Sivic
K
kjkkiji dzpzwpdwp
1
)|()|()|(
![Page 33: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/33.jpg)
)|(maxarg dzpzz
Case #2: Recognition using pLSACase #2: Recognition using pLSA
Slide credit: Josef Sivic
![Page 34: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/34.jpg)
Maximize likelihood of data using EM
Observed counts of word i in document j
M … number of codewords
N … number of images
Case #2: Learning the pLSA parametersCase #2: Learning the pLSA parameters
Slide credit: Josef Sivic
![Page 35: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/35.jpg)
DemoDemo
• Course website
![Page 36: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/36.jpg)
task: face detection – no labelingtask: face detection – no labeling
![Page 37: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/37.jpg)
• Output of crude feature detector– Find edges– Draw points randomly from edge set– Draw from uniform distribution to get scale
Demo: feature detectionDemo: feature detection
![Page 38: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/38.jpg)
Demo: learnt parametersDemo: learnt parameters
Codeword distributionsper theme (topic)
Theme distributionsper image
)|( zwp )|( dzp
• Learning the model: do_plsa(‘config_file_1’)• Evaluate and visualize the model: do_plsa_evaluation(‘config_file_1’)
![Page 39: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/39.jpg)
Demo: recognition examplesDemo: recognition examples
![Page 40: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/40.jpg)
• Performance of each theme
Demo: categorization resultsDemo: categorization results
![Page 41: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/41.jpg)
category modelscategory models(and/or) classifiers(and/or) classifiers
Learning and Learning and RecognitionRecognition
1. Generative method: - graphical models
2. Discriminative method: - SVM
![Page 42: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/42.jpg)
Zebra
Non-zebra
Decisionboundary
Discriminative methods based on ‘bag of words’ representation
![Page 43: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/43.jpg)
Discriminative methods based on ‘bag of words’ representation
• Grauman & Darrell, 2005, 2006:– SVM w/ Pyramid Match kernels
• Others– Csurka, Bray, Dance & Fan, 2004– Serre & Poggio, 2005
![Page 44: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/44.jpg)
Summary: Pyramid match kernel
optimal partial matching between
sets of features
Grauman & Darrell, 2005, Slide credit: Kristen Grauman
![Page 45: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/45.jpg)
Pyramid Match (Grauman & Darrell 2005)
Histogram intersection
Slide credit: Kristen Grauman
![Page 46: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/46.jpg)
Difference in histogram intersections across levels counts number of new pairs matched
matches at this level matches at previous level
Histogram intersection
Pyramid Match (Grauman & Darrell 2005)
Slide credit: Kristen Grauman
![Page 47: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/47.jpg)
Pyramid match kernel
• Weights inversely proportional to bin size
• Normalize kernel values to avoid favoring large sets
measure of difficulty of a match at level i
histogram pyramids
number of newly matched pairs at level i
Slide credit: Kristen Grauman
![Page 48: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/48.jpg)
Example pyramid matchLevel 0
Slide credit: Kristen Grauman
![Page 49: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/49.jpg)
Example pyramid matchLevel 1
Slide credit: Kristen Grauman
![Page 50: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/50.jpg)
Example pyramid matchLevel 2
Slide credit: Kristen Grauman
![Page 51: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/51.jpg)
Example pyramid match
pyramid match
optimal match
Slide credit: Kristen Grauman
![Page 52: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/52.jpg)
Summary: Pyramid match kernel
optimal partial matching between
sets of features
number of new matches at level idifficulty of a match at level i
Slide credit: Kristen Grauman
![Page 53: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/53.jpg)
Object recognition results
• ETH-80 database 8 object classes (Eichhorn and Chapelle 2004)
• Features: – Harris detector– PCA-SIFT descriptor, d=10
Kernel Complexity Recognition rate
Match [Wallraven et al.] 84%
Bhattacharyya affinity [Kondor & Jebara]
85%
Pyramid match 84%
Slide credit: Kristen Grauman
![Page 54: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/54.jpg)
Object recognition results
• Caltech objects database 101 object classes
• Features:– SIFT detector– PCA-SIFT descriptor, d=10
• 30 training images / class• 43% recognition rate (1% chance performance)• 0.002 seconds per match
Slide credit: Kristen Grauman
![Page 55: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/55.jpg)
categorycategorydecisiondecision
learninglearning
feature detection& representation
codewords dictionarycodewords dictionary
image representation
category modelscategory models(and/or) classifiers(and/or) classifiers
recognitionrecognition
![Page 56: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/56.jpg)
What about spatial info?What about spatial info?
?
![Page 57: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/57.jpg)
What about spatial info?What about spatial info?
• Feature level– Spatial influence through correlogram features:
Savarese, Winn and Criminisi, CVPR 2006
![Page 58: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/58.jpg)
What about spatial info?What about spatial info?
• Feature level• Generative models
– Sudderth, Torralba, Freeman & Willsky, 2005, 2006– Niebles & Fei-Fei, CVPR 2007
![Page 59: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/59.jpg)
What about spatial info?What about spatial info?
• Feature level• Generative models
– Sudderth, Torralba, Freeman & Willsky, 2005, 2006– Niebles & Fei-Fei, CVPR 2007
P3
P1 P2
P4
BgImage
w
![Page 60: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/60.jpg)
What about spatial info?What about spatial info?
• Feature level• Generative models• Discriminative methods
– Lazebnik, Schmid & Ponce, 2006
![Page 61: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/61.jpg)
Invariance issuesInvariance issues• Scale and rotation
– Implicit– Detectors and descriptors
Kadir and Brady. 2003
![Page 62: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/62.jpg)
• Scale and rotation
• Occlusion– Implicit in the models– Codeword distribution: small variations– (In theory) Theme (z) distribution: different
occlusion patterns
Invariance issuesInvariance issues
![Page 63: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/63.jpg)
• Scale and rotation
• Occlusion
• Translation– Encode (relative) location information
• Sudderth, Torralba, Freeman & Willsky, 2005, 2006
• Niebles & Fei-Fei, 2007
Invariance issuesInvariance issues
![Page 64: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/64.jpg)
• Scale and rotation
• Occlusion
• Translation
• View point (in theory)– Codewords: detector
and descriptor– Theme distributions:
different view points
Invariance issuesInvariance issues
Fergus, Fei-Fei, Perona & Zisserman, 2005
![Page 65: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/65.jpg)
Model propertiesModel propertiesOf all the sensory impressions proceeding to the brain, the visual experiences are the dominant ones. Our perception of the world around us is based essentially on the messages that reach the brain from our eyes. For a long time it was thought that the retinal image was transmitted point by point to visual centers in the brain; the cerebral cortex was a movie screen, so to speak, upon which the image in the eye was projected. Through the discoveries of Hubel and Wiesel we now know that behind the origin of the visual perception in the brain there is a considerably more complicated course of events. By following the visual impulses along their path to the various cell layers of the optical cortex, Hubel and Wiesel have been able to demonstrate that the message about the image falling on the retina undergoes a step-wise analysis in a system of nerve cells stored in columns. In this system each cell has its specific function and is responsible for a specific detail in the pattern of the retinal image.
sensory, brain, visual, perception,
retinal, cerebral cortex,eye, cell, optical
nerve, imageHubel, Wiesel
• Intuitive– Analogy to documents
![Page 66: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/66.jpg)
Model propertiesModel properties
• Intuitive• generative models
– Convenient for weakly- or un-supervised, incremental training
– Prior information– Flexibility (e.g. HDP)
Li, Wang & Fei-Fei, CVPR 2007
model
Classification
Dataset Incremental
learning
Sivic, Russell, Efros, Freeman, Zisserman, 2005
![Page 67: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/67.jpg)
Model propertiesModel properties
• Intuitive• generative models• Discriminative method
– Computationally efficient
Grauman et al. CVPR 2005
![Page 68: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/68.jpg)
Model propertiesModel properties
• Intuitive• generative models• Discriminative method• Learning and
recognition relatively fast– Compare to other
methods
![Page 69: Part 1: Bag-of-words models by Li Fei-Fei (Princeton)](https://reader037.vdocuments.net/reader037/viewer/2022102714/56649cbf5503460f94984ccf/html5/thumbnails/69.jpg)
• No rigorous geometric information of the object components
• It’s intuitive to most of us that objects are made of parts – no such information
• Not extensively tested yet for– View point invariance– Scale invariance
• Segmentation and localization unclear
Weakness of the modelWeakness of the model