An efficient implementation of Reid's multiple hypothesis tracking algorithm and its evaluation for the purpose of visual tracking

Cox, Ingemar J.; Hingorani, Sunita L.

doi:10.1109/34.481539

Cited by 580 publications

(398 citation statements)

References 26 publications

(18 reference statements)

Supporting

Mentioning

382

Contrasting

Unclassified

Order By: Relevance

“…This approach that considers future positions of the object in addition to the past has been adopted in some automated tracking algorithms including MHT (e.g. Reid, 1979;Cox and Hingorani, 1996;Blackman, 2004;Lakshmanan and Smith, 2010;Scharenbrouch et al, 2010;Root et al, 2011;Lakshmanan et al, 2013;Miller et al, 2013) and ThOR (Barjenbruch and Houston, 2006;Barjenbruch, 2008;Lahowetz et al, 2010) and is employed in AALTO. Future positions are examined by building a search tree in which each node represents the position of an object along a candidate track, which refers to a possible continuation of the existing track.…”

Section: Trackingmentioning

confidence: 99%

The advanced algorithm for tracking objects (AALTO)

Limpert

Houston

Lock

2015

Meteorological Applications

View full text Add to dashboard Cite

ABSTRACT:The advanced algorithm for the tracking of objects (AALTO) constructs tracks from objects, such as thunderstorms or mesocyclones, detected by multiple weather radars at irregular time intervals. It is important to have high accuracy in tracking thunderstorms to generate skilful forecasts and high-quality climatologies and, fundamentally, to ensure that any derived product from tracks captures only that particular storm, and in its entirety. AALTO incorporates many of the best practices of existing tracking algorithms and techniques employed by meteorologists in constructing tracks. AALTO differs from existing algorithms designed to track meteorological phenomena that manifest in radar data in the following ways: (1) AALTO is designed to track objects from multiple radars, enabling analysis over a larger domain than if a single radar was used; (2) improved tracking is realized through improved initial motion estimates and directional thresholding and (3) AALTO looks both at the track history and at the subsequent possible positions along a track when constructing the best possible tracks, mimicking the approach that would be taken by a human meteorologist. Verification was done using metrics that were objectively determined to distinguish between good and degraded tracks; a description of the approach to determine the appropriate metrics is presented. An overview of the AALTO tracking procedure and an example case are presented in this study.

show abstract

Section: Trackingmentioning

confidence: 99%

The advanced algorithm for tracking objects (AALTO)

Limpert

Houston

Lock

2015

Meteorological Applications

View full text Add to dashboard Cite

show abstract

“…35,47 Statistical methods, for example using probability density functions such as Multiple Hypothesis Trackers and NP-Hard methods have been applied with varying degrees of success. 40,48 Multiple Hypothesis Trackers (MHT) introduce probabilistic knowledge, but can be difficult to apply from matching a variable number of image feature points in global optimization both in space and time, while NP-Hard methods prohibit fast computation. Heuristic methods can be used to identify putative tracks from qualitative descriptions, but suffer the disadvantage of relying critically on the accuracy in the detection stage and easily fail when ambiguities occur.…”

Section: Pinpointing Fluorescently-labelled Moleculesmentioning

confidence: 99%

“…Heuristic methods can be used to identify putative tracks from qualitative descriptions, but suffer the disadvantage of relying critically on the accuracy in the detection stage and easily fail when ambiguities occur. 48,49 To overcome many of the shortcomings of existing tracking methods, my laboratory has developed a new automated multiparticle tracking algorithm based on minimal path optimization, similar to those used previously but extended in application to native cell membranes of living bacterial cells. After detecting candidate particles and linking image feature points frame-byframe, some segmented trajectories are obtained initially.…”

Section: Pinpointing Fluorescently-labelled Moleculesmentioning

confidence: 99%

Analytical tools for single-molecule fluorescence imaging in cellulo

Leake

2014

Phys. Chem. Chem. Phys.

View full text Add to dashboard Cite

Recent technological advances in cutting-edge ultrasensitive fluorescence microscopy have allowed single-molecule imaging experiments in living cells across all three domains of life to become commonplace. Single-molecule live-cell data is typically obtained in a low signal-to-noise ratio (SNR) regime sometimes only marginally in excess of 1, in which a combination of detector shot noise, suboptimal probe photophysics, native cell autofluorescence and intrinsically underlying stochasticity of molecules result in highly noisy datasets for which underlying true molecular behaviour is nontrivial to discern. The ability to elucidate real molecular phenomena is essential in relating experimental single-molecule observations to both the biological system under study as well as offering insight into the fine details of the physical and chemical environments of the living cell. To confront this problem of faithful signal extraction and analysis in a noise-dominated regime, the 'needle in a haystack' challenge, such experiments benefit enormously from a suite of objective, automated, high-throughput analysis tools that can home in on the underlying 'molecular signature' and generate meaningful statistics across a large population of individual cells and molecules. Here, I discuss the development and application of several analytical methods applied to real case studies, including objective methods of segmenting cellular images from light microscopy data, tools to robustly localize and track single fluorescently-labelled molecules, algorithms to objectively interpret molecular mobility, analysis protocols to reliably estimate molecular stoichiometry and turnover, and methods to objectively render distributions of molecular parameters.

show abstract

“…multi-hypothesis detection problems [9,10,7,8], without restricting to [4]. Then we derive continuous valued features (e.g., {Dist j , P robDecay j }) more precisely describing the underlying "candidate-ICV" associations, which permits further statistical analysis and classification, converting from binary detections.…”

Section: Discussionmentioning

confidence: 99%

“…1. For robustness, all steps of this detection pipeline leverage and keep multiple hypotheses (as a set of 3D boxes) for the next level until the last stage, which is in the same spirit of robust object tracking using multiple hypotheses [7], sequential Monte Carlo or particle filtering [8].…”

Section: Introductionmentioning

confidence: 99%

Correcting Misalignment of Automatic 3D Detection by Classification: Ileo-Cecal Valve False Positive Reduction in CT Colonography

Lü

Wolf

et al. 2011

Medical Computer Vision. Recognition Techniques and Applications in Medical Imaging

View full text Add to dashboard Cite

Abstract. Ileo-Cecal Valve (ICV) is an important small soft organ which appears in human abdomen CT scans and connects colon and small intestine. Automated detection of ICV is of great clinical value for removing false positive (FP) findings in computer aided diagnosis (CAD) of colon cancers using CT colongraphy (CTC) [1][2][3]. However full 3D object detection, especially for small objects with large shape and pose variations as ICV, is very challenging. The final spatial detection accuracy often trades for robustness to find instances under variable conditions [4].In this paper, we describe two significant post-parsing processes after the normal procedure of object (e.g., ICV) detection [4], to probabilistically interpret multiple hypotheses detections. It achieves nearly 300% performance improvement on (polyp detection) FP removal rate of [4], with about 1% extra computional overhead. First, a new multiple detection spatial-fusion method utilizes the initial single detection as an anchor identity and iteratively integrates other "trustful" detections by maximizing their spatial gains (if included) in a linkage. The ICV detection output is thus a set of N spatially connected boxes instead of a single box as top candidate, which shows to correct 3D detection misalignment inaccuracy. Next, we infer the spatial relationship between CAD generated polyp candidates and the detected ICV bounding boxes in 3D volume, and convert as a set of continuous valued, ICV-association features per candidate which allows further statistical analysis and classification for more rigorous false positive deduction in colon CAD.Based on our annotated 116 training cases, the spatial coverage ratio between the new N-box ICV detection and annotation is improved by 13.0% (N=2) and 19.6% (N=3) respectively. An evaluation on large scale datasets of total ∼ 1400 CTC volumes, with different tagging preparations, reports average 5.1 FP candidates are removed at CandidateGeneration stage per scan; and the final CAD system mean FP rate drops from 2.2 to 1.82 per volume, without affecting the sensitivity.

show abstract

An efficient implementation of Reid's multiple hypothesis tracking algorithm and its evaluation for the purpose of visual tracking

Cited by 580 publications

References 26 publications

The advanced algorithm for tracking objects (AALTO)

The advanced algorithm for tracking objects (AALTO)

Analytical tools for single-molecule fluorescence imaging in cellulo

Correcting Misalignment of Automatic 3D Detection by Classification: Ileo-Cecal Valve False Positive Reduction in CT Colonography

Contact Info

Product

Resources

About