People Detection benchmark repository
Results
Author: Pedro F. Felzenszwalb.
Contact university/company: University of Chicago.
Contact email: pff@cs.uchicago.edu
Method name: Discriminatively Trained Deformable Part Model(DTDP) .
Description and pocessing time: The DTDP detector is based on exhaustive search and a part-based person model. The DTDP approach is implemented with Matlab and the computational cost is around 2 seconds per frame with 352x288 images (there is a faster implementation in OpenCV that runs around 1 second per frame). The tests have been performed on a Pentium IV with a CPU frequency of 2.4 GHz and 3GB RAM.
Parameters: Author default parameters here (voc-release4).
You can download the people detection results of this method by clicking here.
Results for each video sequence:
Video |
DTDP |
1 |
96.7 |
2 |
66.3 |
3 |
69.7 |
4 |
13.2 |
5 |
85.1 |
6 |
74.9 |
7 |
11.0 |
8 |
0.0 |
9 |
90.4 |
10 |
0.7 |
11 |
11.4 |
12 |
92.4 |
13 |
80.2 |
14 |
44.1 |
15 |
67.9 |
16 |
83.3 |
Average AUC |
55.4 |
Ranking |
3.00 |
Results for each background complexity:
Background complexity |
DTDP |
Baseline |
77.6 |
Dynamic Background |
49.2 |
Camera Jitter |
42.9 |
Intermittent Object Motion |
25.6 |
Shadow |
73.6 |
Average AUC |
53.8 |
Ranking |
2.20 |
Results for each classification complexity:
Classification complexity |
DTDP |
Low |
84.9 |
Medium |
74.4 |
High |
13.4 |
Average AUC |
57.6 |
Ranking |
2.00 |
