Pedestrian Recognition

1
Pedestrian Recognition

• Machine Perception and Modeling of Human Behavior
• Manfred Lau

2
Pedestrian Recognition

• Oren, Papageorgiou, Sinha, Osuna, Poggio.
  Pedestrian Detection Using Wavelet Templates.
  CVPR 1997.
• Papageorgiou, and Poggio. Trainable Pedestrian
  Detection. International Conference on Image
  Processing 1999.

3
Motivation

• Recognition system inside vehicles
• Valerie detect and greet those who stop in
  front of the booth

4
Overview
Positive samples
Negative samples
Classifier
5
Wavelet Template
-1
1
vertical wavelet
Average of many samples

• Compute coefficient for each RGB channel and take
  largest absolute value
• Vertical wavelet identifies vertical color
  differences

6
Wavelet Template
Average of many samples
7
Features

• Each image is one instance with 1326 features and
  one classification
• Same thing for negative samples

8
Test case

• 282 positive samples, 236 negative samples for
  training
• 20 positives and 20 negatives for testing

Some Positive Samples
9
Some negative samples
10
Results
Nearest neighbor classifier 95 accuracy
Decision tree classifier 90 accuracy
2 false positives
3 false positives, 1 false negative
11
10-fold cross validation

• Test case 302 positives, 256 negatives
• Nearest neighbor ? 94.27
• 30 false positives, 2 false negatives
• Decision tree ? 86.74
• 47 false positives, 27 false negatives

12
Incremental bootstrapping

• Use nearest neighbor
• But problem with many false positives

13
Incremental bootstrapping

• Took database of 558 total samples
• After bootstrapping, 656 total samples

14
Bootstrapping
15
Result

• A completely new
• test image
• Before bootstrapping
• 85.06 accurate, 65 false pos, 0 false neg
• After bootstrapping
• 90.11 accurate, 43 false pos, 0 false neg

16
Result

• Another new
• test image
• Before bootstrapping
• 75.86 accurate, 100 false pos, 5 false neg
• After bootstrapping
• 81.15 accurate, 77 false pos, 5 false neg

17
Splitted up into 560 images, about 30 classified
as positive
Some false positives
true positives
18
Results
19
Less features

• Take average coefficients across many positive
  samples
• Pick those features that are darkest/lightest ?
  can use much less than 1326 features, for faster
  classification

20
Conclusions

• Can detect positive samples well, but many false
  positives
• Bootstrapping on more and more new images will
  decrease false positives (Im not doing enough of
  this)

21
Limitations

• Recognize only template, other objects may be
  similar
• Difficult to define what is a negative sample
• What if pedestrians are partially occluded?