Introduction to Recognition

1
Introduction to Recognition
CS4670/5670 Intro to Computer Vision
Noah Snavely
2
What do we mean by object recognition?
Next 15 slides adapted from Li, Fergus,
Torralbas excellent short course on category and
object recognition
3
Verification is that a lamp?
4
Detection are there people?
5
Identification is that Potala Palace?
6
Object categorization
mountain
tree
building
banner
street lamp
vendor
people
7
Scene and context categorization

• outdoor
• city

8
Object recognitionIs it really so hard?
Output of normalized correlation
9
Object recognitionIs it really so hard?
Pretty much garbage Simple template matching is
not going to make it
10
Object recognitionIs it really so hard?
A popular method is that of template matching,
by point to point correlation of a model pattern
with the image pattern. These techniques are
inadequate for three-dimensional scene analysis
for many reasons, such as occlusion, changes in
viewing angle, and articulation of parts.
Nivatia Binford, 1977.
11
Why not use SIFT matching for everything?

• Works well for object instances
• Not great for generic object categories

12
Applications Computational photography
13
Applications Assisted driving
Pedestrian and car detection
Lane detection

• Collision warning systems with adaptive cruise
  control,
• Lane departure warning systems,
• Rear object detection systems,

14
Applications image search
15
How do human do recognition?

• We dont completely know yet
• But we have some experimental observations.

16
Observation 1

• We can recognize familiar faces even in
  low-resolution images

17
Observation 2
Kevin Costner
Jim Carrey

• High frequency information is not enough

18
What is the single most important facial features
for recognition?
19
Observation 4

• Image Warping is OK

20
The list goes on

• http//web.mit.edu/bcs/sinha/papers/19results_sinh
  a_etal.pdf

21
Questions?