event thumbnail image
Machine Learning over Text & Images - Autumn School

Generative Models for Visual Objects and Object Recognition via Bayesian Inference

author: Fei-Fei Li, Princeton University
You might be experiencing some problems with Your Video player.
Slides
0:00 Generative Models for Visual Objects and Object Recognition via Bayesian Inference
0:47 outline
1:16 picture 1
1:40 Plato said…
2:49 picture 2
3:13 How many object categories are there?
4:35 So what does object recognition involve?
4:45 Verification: is that bus?
5:00 Detection: are there cars?
5:24 Identification: is this a picture of Mao?
5:51 Object categorization
6:26 Scene and context categorization
7:00 Challenges 1: view point variation
7:49 Challenges 2: illumination
8:24 Challenges 3: occlusion
9:10 Challenges 4: scale
9:35 Challenges 5: deformation
9:59 Challenges 6: background clutter
10:23 History: single object recognitionv - part 1
11:15 History: single object recognition - part 2
11:37 Challenges 7: intra-class variation
12:28 History: early object categorizationn - part 1
13:03 History: early object categorization - part 2
13:30 picture 3
14:11 objects
14:37 picture 4
15:22 Scenes, Objects, and Parts
15:36 Object categorization: the statistical viewpoint - part 1
17:51 Object categorization: the statistical viewpoint - part 2
18:44 Discriminative
18:59 Generative
20:08 Three main issues
21:09 Representation - part 1
21:28 Representation - part 2
22:01 Representation - part 3
22:26 Representation - part 4
23:17 Representation - part 5
23:33 Learning - part 1
24:05 Learning - part 2
24:30 Learning - part 3
25:23 Learning - part 4
25:45 Learning - part 5
26:15 Learning - part 6
26:29 Recognition
27:12 Bag-of-words models
27:59 Related works
28:10 Object - bag of words
28:37 Analogy to documents
29:43 picture 5
30:07 learning - recognition
30:56 Representation
31:07 1.Feature detection and representation - part 1
31:10 1.Feature detection and representation - part 2
31:38 1.Feature detection and representation - part 3
32:00 1.Feature detection and representation - part 4
32:02 1.Feature detection and representation - part 5
32:52 1.Feature detection and representation - part 6
32:59 2. Codewords dictionary formation - part 1
33:04 2. Codewords dictionary formation - part 2
34:19 2. Codewords dictionary formation - part 3
34:25 1.Feature detection and representation - part 6
35:11 2. Codewords dictionary formation - part 2
35:31 2. Codewords dictionary formation - part 3
36:00 Image patch examples of codewordsImage codewords
36:06 3. Image representation
36:38 Representation
36:46 Learning and Recognitio
36:54 2 case studies
37:29 First, some notations
38:46 Case #1: the Na Naïve ve Bayes model
40:10 picture 6
40:19 Confusion matrix
40:25 Case #2: Hierarchical Bayesian text models
40:55 Case #2: Hierarchical Bayesian text models - part 1
41:30 Case #2: Hierarchical Bayesian text models - part 2
41:37 Case #2: Hierarchical Bayesian text models - part 1
41:53 Case #2: Hierarchical Bayesian text models - part 2
43:09 Case #2: the pLSA model - part 1
43:14 Case #2: the pLSA model - part 2
44:06 Case #2: Recognition using pLSA
44:20 Case #2: Learning the pLSA parameters
44:57 Learning and Recognition
45:05 Invariance issues - part 1
45:52 Invariance issues - part 1
46:06 Invariance issues - part 2
46:48 Invariance issues - part 3
46:58 Invariance issues - part 4
47:18 Model properties - part 1
47:27 Model properties - part 2
48:00 Model properties - part 3
48:09 Weakness of the mode
48:56 part-based models
49:03 Problem with bag-of-words
49:51 Overview of section
50:12 Representation
50:15 Model: Parts and Structure
50:51 Representation
51:07 Example scheme
52:27 Sparse representation
53:28 History of Idea
53:41 The correspondence problem
54:32 Connectivity of parts - part 1
55:02 Connectivity of parts - part 2
55:05 Different graph structures
55:40 Regions or pixels
56:11 How to model location?
57:13 Explicit shape model
57:53 Shape
58:35 Euclidean & Affine Shape
58:43 Other invariance methods
58:49 Representation of appearance - part 1
59:06 Representation of appearance - part 2
59:08 Representation of occlusion
59:11 Representation of background clutter
59:46 Learning
59:48 Learning situations
60:00 picture 6
60:03 Learning using EM
61:08 Example scheme, using EM for maximum likelihood learning
61:31 Priors
63:14 Learning Shape & Appearance simultaneously
63:18 Number of training examples
63:24 Recognition
63:36 What task?
63:52 Efficient search methods
64:05 Parts and Structure demo
64:11 Demo images
64:14 Online resources
65:04 List properties of ideal recognition system

Lecture rating

People found this lecture:
Worth seeing
because it is:
 Valuable and informative
Well presented
Easily understandable
Acceptably recorded
You need to login to cast your vote.

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.

Related content

Link this page

Would you like to put a link to this lecture on your homepage?
Go ahead! Copy the HTML snippet !

Reviews and comments:

Comment1 Durk Kingma, October 27, 2007 at 10:12 p.m.:

Awesome lecture! A very clear overview of image learning recognition techniques. It's a shame its only one hour, I hope more of her lectures will become available...

Comment2 Jeff Henderson, June 17, 2008 at 4:40 a.m.:

Well-organized lecture that frames the problem of generalized object recognition very well.

Comment3 alex, June 21, 2008 at 11:26 a.m.:

very good, easy understand and comprehensive introduction.

Write your own review or comment: