1. Recognizing and Learning Object categories
Summary

2. Summary Methods reviewed here Resources online Bag of words
Parts and structure
Discriminative methods
Combined Segmentation and recognition
Resources online
Slides
Code
Links to datasets

3. List properties of ideal recognition system
Representation
1000’s categories,
Handle all invariances (occlusions, view point, …)
Explain as many pixels as possible (or answer as many questions as you can about the object)
fast, robust
Learning
Handle all degrees of supervision
Incremental learning
Few training images …

4. Online resources

5. Links to datasets
The next tables summarize some of the available datasets for training and testing object detection and recognition algorithms. These lists are far from exhaustive.
Databases for object localization
CMU/MIT frontal faces
vasc.ri.cmu.edu/idb/html/face/frontal_images
cbcl.mit.edu/software-datasets/FaceData2.html
Patches
Frontal faces
Graz-02 Database
Segmentation masks
Bikes, cars, people
UIUC Image Database
l2r.cs.uiuc.edu/~cogcomp/Data/Car/
Bounding boxes
Cars
TU Darmstadt Database
Motorbikes, cars, cows
LabelMe dataset
people.csail.mit.edu/brussell/research/LabelMe/intro.html
Polygonal boundary
>500 Categories
Databases for object recognition
Caltech 101
Segmentation masks
101 categories
COIL-100
www1.cs.columbia.edu/CAVE/research/softlib/coil-100.html
Patches
100 instances
NORB
Bounding box
50 toys
On-line annotation tools
ESP game
Global image descriptions
Web images
LabelMe
people.csail.mit.edu/brussell/research/LabelMe/intro.html
Polygonal boundary
High resolution images
Collections
PASCAL
Segmentation, boxes
various

6. LabelMe Dataset There are about 19,500 labelled objects
This presentation is about LabelMe: the online open annotation tool. This is a joint project with Bryan, Antonio, and Bill. You can find the page by Google searching "LabelMe MIT".
Google search: LabelMe MIT

7. LabelMe Screen Shot
To give you a flavor of the tool, here is a screen shot of the tool. The tool fits entirely in the browser and is designed with simplicity in mind. To annotate an object, the user simply clicks points around the object. Upon completion, a bubble appears and queries the user to enter the label of the object. The user can delete existing labels or edit them by using the mouse to select the label of interest. The right side of the tool lists the current labels and gives a brief explanation of the task. The user can label a new image by clicking on the "Show New Image" button at the top. The user can also see some statistics that have been gathered on the annotation set.

8. Matlab toolbox
LMquery (database, 'object.name', 'car,building,road,tree')

9. Toolbox LMquery (database, 'object.name', 'car,building,road,tree')
LMcookdatabase (database, 'objectname', 'screen', …
'objectsize', [64 64], 'objectlocation', 'original','maximagesize', [ ])

10. Example Annotations
Here are some example annotations of some office scenes that have been produced by the annotation tool. These figures were generated using the MATLAB tool to query various objects in the dataset.