An opposition to WindowScanning Approaches in Computer Vision

1
An opposition to Window-Scanning Approaches in
Computer Vision

• Presented by Tomasz Malisiewicz
• March 6, 2006
• Advanced Perception _at_ The Robotics Institute

2
(No Transcript)
3
2 Problems

• Does scanning windows across an image work?
• What types of objects does it work for?

4
What are window-scanning approaches missing?
5
Quick Question What is this?
6
What is context?

• Any data or meta-data not directly produced by
  the presence of an object
• Nearby image data

7
What is context?

• Any data or meta-data not directly produced by
  the presence of an object
• Nearby image data
• Scene information

Context
Context
8
What is context?

• Any data or meta-data not directly produced by
  the presence of an object
• Nearby image data
• Scene information
• Presence, locations of other objects

Tree
9
Clues for Function

• What is this?

10
Clues for Function

• What is this?
• Now can you tell?

11
Low-Res Scenes

• What is this?

12
Low-Res Scenes

• What is this?
• Now can you tell?

13
More Low-Res

• What are these blobs?

14
More Low-Res

• The same pixels! (a car)

15
Why is context useful?

• Objects defined at least partially by function
• Trees grow in ground
• Birds can fly (usually)
• Door knobs help open doors

16
Why is context useful?

• Objects defined at least partially by function
• Context gives clues about function
• Not rooted into the ground ? not tree
• Object in sky ? cloud, bird, UFO, plane,
  superman
• Door knobs always on doors

17
Why is context useful?

• Objects defined at least partially by function
• Context gives clues about function
• Objects like some scenes better than others
• Toilets like bathrooms
• Fish like water

18
Why is context useful?

• Objects defined at least partially by function
• Context gives clues about function
• Objects like some scenes better than others
• Many objects are used together and, thus, often
  appear together
• Kettle and stove
• Keyboard and monitor

19
The other problem

• What types of objects does it work for?

Assuming we can just directly avoid the first
problem
20

• Our goal is to develop a system that detects and
  recognizes many kinds of objects in photographs
  and video including everyday office objects, text
  captions in video, and various structures in
  biomedical imagery. Schneiderman and Kanade
  from Object Detection Using the Statistics of
  Parts

However, such approaches seem unlikely to scale
up to the detection of hundreds or thousands of
different object classes because each classifier
is trained and run independently. Torralba
and Murphy and Freeman from Sharing features
efficient boosting procedures for multiclass
object detection
How many different classifiers must one
construct? A different classifier for each
object? A different classifier for each pose of
an object? How many poses do we need per object?
21
Too many windows

• Now imagine scanning a window and applying 100K
  independent classifiers at each window

22
Conclusion

• Without context, we cant find all things we want
  to find. We need context to help constrain the
  search for objects.
• With independent classifiers per object (and per
  pose), we cant detect a large number of objects.
  Should cow detectors and a horse detectors be
  built independently? Think along the lines of a
  horse and a cow are types of animals that often
  occur in similar contexts.
• Remember that complex and deformable objects
  would require many poses if are to adhere to the
  window-based classifier paradigm.

23
Thank you.
Pascal 2006 Visual Challenge Image