116. State-of-the-art : Deformable Part Model
[Felzenszwalb,PAMI 10]
• 人全体を捉えるルートフィルタとパーツフィルタの
位置関係を自動獲得
116
ルートフィルタ パーツフィルタ
パーツフィルタの
位置関係
117. DPMによる誤検出例
117
Fig. 6. Qualitative detection results on the Ca
row shows typical false positives.
laborative Research Lab.
REFERENCES
[1] A. Bar-Hillel, D. Levi, E. Krupka, and C. G
synthesis for human detection. In ECCV,
[2] H. Cho, P. Rybski, and W. Zhang. Vision-
tracking using a deformable part model an
2010.
on the Caltech testset. The first and second row shows correct pedestrian detections in various scenarios. The third
ENCES
a, and C. Goldberg. Part-based feature
In ECCV, 2010.
g. Vision-based bicycle detection and
model and an ekf algorithm. In ITSC,
[13] P. F. Felzenszwalb, R. B. Girshick, and D. McAllester.
Discriminatively trained deformable part models, release 4.
http://people.cs.uchicago.edu/⇠pff/latent-release4/.
[14] D. M. Gavrila and S. Munder. Multi-cue pedestrian detection and
tracking from a moving vehicle. International Journal of Computer
Vision, 73:41–59, 2007.
[15] D. Gero´nimo, A. Lo´pez, and T. G. A. Sappa. Survey of pedestrian
detection for advanced driver assistance systems. IEEE Transaction
on Pattern Analysis and Machine Intelligence, 32, 2010.
an detections in various scenarios. The third
R. B. Girshick, and D. McAllester.
d deformable part models, release 4.
edu/⇠pff/latent-release4/.
Munder. Multi-cue pedestrian detection and
vehicle. International Journal of Computer
and T. G. A. Sappa. Survey of pedestrian
river assistance systems. IEEE Transaction
Machine Intelligence, 32, 2010.
HOG(特徴抽出過程)の限界
(Real-time Pedestrian Detection with Deformable Part Models, IV2012)
134. 参考文献1
•1. Haar-like特徴量と顔検出
- [Viola & Jones, CVPR 01] Viola, P and Jones, M, Rapid object detection
using a boosted cascade of simple features , CVPR,vol.1,pp.511-518,
(2001).
• 2. AdaBoost
- [Freund 97] Y, Freund and R, E. Schapire, A decisiontheoretic
generalization of on-line learning and an application to boosting , Journal of
Computer and System Sciences, No. 1, Vol. 55, pp. 119-139,(1997).
• 3. HOG特徴量
- [Dalal 05] Dalal. N, Triggs. B, Histograms of Oriented Gradients for Human
Detection , IEEE CVPR, pp. 886-893 (2005).
•4. Real AdaBoost
- [Schapire 99] R. E. Schapire, Y. Singer, Improved Boosting Algorithms
Using Confidence-rated Predictions , Machine Learning, No. 37, pp.
297-336, (1999).
134
135. •5. 人検出
- [Comaniciu 02] D. Comaniciu, P. Meer, Mean Shift: A Robust Approach
toward Feature Space Analysis , IEEE PAMI, vol. 24, No. 5, pp. 603-619,
(2002).
•6. 第3世代の特徴量
- [Rowley, PAMI 98] H.A. Rowley, S. Baluja, and T. Kanade, Neural network-
based face detection ,IEEE Transactions on PAMI,vol.20,pp.23-28,(1998).
- [Mita et al., PAMI 08] T. Mita, T. Kaneko, B. Stenger, and O.
Hori:``Discriminative Feature Co-occurrence Selection for Object
Detection,'' Pattern Analysis and Machine Intelligence, Vol.30, no.7, pp.
1257-1269(2008)
- [三井, 山内, 藤吉 SSII 08] 三井 相和,山内 悠嗣,藤吉 弘亘, アピアランスと時空
間特徴を用いたJoint特徴による人検出 ,電気関係学会東海支部連合大会,pp.O-
135,(2008).
- E.Hinton, G.,Osindero,S.and Teh, Y.-W.:A fast learning algorithm for deep
belief nets, Neural Computation, Vol.18, pp.1527-1544(2006).
135
参考文献2
136. - E.Hinton, G.,Osindero,S.and Teh, Y.-W.:A fast learning algorithm for deep
belief nets, Neural Computation, Vol.18, pp.1527-1544(2006).
- [Felzenszwalb,PAMI 10] P. Felzenszwalb, R. Girshick, D. McAllester and D.
Ramanan, Object Detection with Discriminatively Trained Part Based
Models ,IEEE Transactions on PAMI,vol.32,no.9,pp.1627-1645,(2010).
- [Felzenszwalb,CVPR 10] P. Felzenszwalb, R. Girshick and D. McAllester,
Cascade Object Detection with Deformable Part Models, IEEE CVPR, pp.
2241‒2248, 2010.
- [Ott,CVPR 10] P. Ott and M. Everingham, Shared Parts for Deformable
Part-Based Models, IEEE CVPR, 2010.
- [Girshick,NIPS 11] R. Girshick, P. Felzenszwalb and D. McAllester, Object
Detection with Grammar Models , NIPS, 2011
136
参考文献3