Performance evaluation of crowd image analysis using the PETS2009 dataset

Ferryman, James; Ellis, Anna-Louise

doi:10.1016/j.patrec.2014.01.005

Cited by 34 publications

(19 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Three cases are considered: Non-regularized histograms, regularized histograms and the incomplete information filling heuristic of 3.6. We test our algorithm on three datasets and we evaluate the performance of the resulting tracking using the VACE and CLEAR metrics [12].…”

Section: Methodsmentioning

confidence: 99%

“…4 and 5, we apply the learning scheme and use the priors in two scenarios: First, under a particle filter approach, where the priors are used in the filter proposal distribution; then, under a graph-based methodology, where the priors are used to weight the graph edges. We evaluate our approach in two challenging datasets, under six standard metrics [12], and compare our results against other proposals (Sect. 6).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Motion priors for multiple target visual tracking

Madrigal

Hayet

Rivera

2015

Machine Vision and Applications

View full text Add to dashboard Cite

This article describes an original strategy for enhancing current state-of-the-art trackers through the use of motion priors, built as data-driven probabilistic motion models for moving targets. Our priors have a simple form and can replace advantageously more traditional models, such as the constant velocity or constant acceleration models, that are of common use in visual tracking systems, but that are also prone to fail in handling critical scene-related constraints on the targets motion. These priors are learned based on local motion observed in the video stream(s) and, given that the obtained representation may be incomplete and noisy, we regularize it in a second phase. The hybrid discretecontinuous motion priors are then used within two classical target tracking approaches: (1) as a sampling distribution in a particle filter framework and (2) as a weighting prior in a detection-based framework. For both tracking schemes, we present promising results with our motion prior approach, on classical benchmark datasets from the visual surveillance tracking literature.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Motion priors for multiple target visual tracking

Madrigal

Hayet

Rivera

2015

Machine Vision and Applications

View full text Add to dashboard Cite

show abstract

“…We benchmarked our algorithm on PETS '09 dataset [46]. This video is filmed to be a reference in object tracking and it is used in many approaches.…”

Section: A Datasetmentioning

confidence: 99%

A Novel Object Position Coding for Multi-Object Tracking using Sparse Representation

Elbahri¹,

Kpalma²,

Taleb³

et al. 2015

IJIGSP

View full text Add to dashboard Cite

Abstract-Multi-object tracking is a challenging task, especially when the persistence of the identity of objects is required. In this paper, we propose an approach based on the detection and the recognition. To detect the moving objects, a background subtraction is employed. To solve the recognition problem, a classification system based on sparse representation is used. With an online dictionary learning, each detected object is classified according to the obtained sparse solution. Each column of the used dictionary contains a descriptor representing an object. Our main contribution is the representation of the moving object with a descriptor derived from a novel representation of its 2-D position and a histogram-based feature, improved by using the silhouette of this object. Experimental results show that the approach proposed for describing moving objects, combined with the classification system based on sparse representation provides a robust multi-object tracker in videos involving occlusions and illumination changes.

show abstract

“…However, this system requires relatively stable lighting conditions (face detection is very computationally-intensive and extremely challenging in uncontrolled environments [12,11]). Other techniques include those based on separating background and foreground (moving) objects [17], extracting features from segmented frames [3], and many others [1,20,23,21,4,19,5,6].…”

Section: Introductionmentioning

confidence: 99%

Real-Time People Counting from Depth Images

Nalepa

Szymanek²,

Kawulok

2015

Beyond Databases, Architectures and Structures

View full text Add to dashboard Cite

Abstract. In this paper, we propose a real-time algorithm for counting people from depth image sequences acquired using the Kinect sensor. Counting people in public vehicles became a vital research topic. Information on the passenger flow plays a pivotal role in transportation databases. It helps the transport operators to optimize their operational costs, providing that the data are acquired automatically and with sufficient accuracy. We show that our algorithm is accurate and fast as it allows 16 frames per second to be processed. Thus, it can be used either in real-time to process traffic information on the fly, or in the batch mode for analyzing very large databases of previously acquired image data.

show abstract

Performance evaluation of crowd image analysis using the PETS2009 dataset

Cited by 34 publications

References 19 publications

Motion priors for multiple target visual tracking

Motion priors for multiple target visual tracking

A Novel Object Position Coding for Multi-Object Tracking using Sparse Representation

Real-Time People Counting from Depth Images

Contact Info

Product

Resources

About