Multiple Hypothesis Tracking
Particle filtering was just a convenient way of doing MHT(multi-hypothesis tracking )
'Spatially Indexed Clustering for Scalable Tracking of Remotely Sensed Drift Ice', 2017 [코드-Python]
C. Kim, "Multiple Hypothesis Tracking Revisited", ICCV 2015 , [코드-Python]
MULTIPLE OBJECT TRACKING WITH MHT : blog 2018
Multiple Hypothesis Tracking: python,
'Spatially Indexed Clustering for Scalable Tracking of Remotely Sensed Drift IceIEEE2017Multi Hypothesis Tracking: by Carsten Haubold, 2015, Python
pyMHT : Python
Multiple Hypothesis Tracking Revisited: Jim Rehg
LIDAR-BASED MULTI-OBJECT TRACKING SYSTEM WITH DYNAMIC MODELING: 2012, 석사 학위 논문
현 (하나의)프레임에서 바로 할당작업을 진행하는 방식과는 달리, MHT는 불확실한 상황이 되면 할당작업을 미루고 여러개의 alternative data association hypotheses를 생성 한다. Instead of considering just the information of one frame, Multiple Hypothesis Tracking (MHT) is a multi-scan deferred decision logic tracking algorithm that keeps multiple alternative data association hypotheses whenever the ambiguity situation occurs.
JPDA에서 이러한 가설치(hypotheses)들을 합치는 방식 대비, MHT에서는 가설치들이 전파되어서 가장 높은 posteriors를 가지는 값만 반환된다. Rather than combining these hypotheses in JPDA method, all the hypotheses are propagated and the one with highest posteriors is returned as solution.
모든 track들이 유지 되기 때문에 모션모델이 불확실한곳에서 유용한다. Since all potential tracks are maintained and updated, this method is very useful when the target motion model is unpredictable.
단점으로 시간이 지남에 따라 유지해야 하는 가설치가 기하 급수적으로 증가하긴 하지만 아직많이 사용되고 있다. [26,28,63]에서는 가설치 삭제 방안도 제시 되었다. Although the number of hypothesis may grow exponentially over time, MHT is still applied in many multi-object tracking systems, supplementary with hypothesis deleting mechanism [26, 28, 63]. Wang, et al. also implemented this method on Navlab for real-time road tests (see the previous Figure 2.9).
Effective Data Association Algorithms for Multitarget Tracking
2.3.2 Multiple Hypothesis Testing
In the Multiple Hypothesis Testing (MHT) [19][78] approach a hypothesis will be generated and tested with the received measurements in the current scan or frame.
For a given measurement the hypotheses could be the measurement is originated from one of initialized tracks, or is originated from a new target or is a false alarm [18].
A Bayesian approach will be used to compute the probabilities of each hypothesis.
The valid hypotheses derived from sequences of measurements are evaluated and propagated over time, each of them generating a set of new hypotheses at every sample time k.
단점 : The major drawback in implementing the MHT algorithm for practical applications is the exponential growth in the number of the assignment hypotheses as time of a scan and number of measurement increases.
This leads to the development of several hypothesis pruning, hypothesis merging and gating techniques [18][29].