Hybrid 3D–2D human tracking in a top view

Migniot, Cyrille; Ababsa, Fakhreddine

doi:10.1007/s11554-014-0429-7

Cited by 23 publications

(30 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In particular, the head's points, selected by Equation (6), are deleted. P C hd (t) − J hd (t − 1) < th hd (6) P C hd (t) represents the points of the head in the current frame, J hd is the head joint calculated from the mean of all the elements in P C hd . th hd is a distance threshold used to find the head points at frame t, based on the position of J hd in the previous frame.…”

Section: Growing Neural Gas Networkmentioning

confidence: 99%

“…Consequently, the network will be parallel to the y-axis, irrespective of how the person is oriented relative to the camera. The triangle highlights the joint J hd calculated with Equation (6). The shoulder joints are selected respectively as the left/right node to J hd with the lowest y coordinate from all the nodes in the network.…”

Section: Growing Neural Gas Networkmentioning

confidence: 99%

“…Using this configuration, it is not possible to exploit the skeleton algorithm provided by Microsoft [5] to track people's movements, so other solutions, which directly use the depth stream, are necessary. A tracking systems based on two particle filters has been proposed by Migniot et al [6]. The same idea is used by Bednařík et al in [7], with good results in terms of performance on a dataset of 10 people.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Performance analysis of self-organising neural networks tracking algorithms for intake monitoring using kinect

Gasparrini¹,

Cippitelli²,

Gambi³

et al. 2015

IET International Conference on Technologies for Active and Assisted Living (TechAAL)

View full text Add to dashboard Cite

The analysis of intake behaviour is a key factor to understand the health condition of a subject, such as elderly or people affected by diet-related disorders. The technology can be exploited for this purpose to promptly identify anomalous situations. To this end, the point cloud, provided by a depth camera placed on the ceiling in top-down view, is used as input to three self-organising algorithms. The output are three different models that represent the monitored person during intake activities. Starting from these models, the nodes representing the head and the hands are selected. They are useful to identify most of the actions performed by the person while having a meal. In the experimental section, the positions of these nodes are compared with a ground truth and the performance of the proposed algorithms are evaluated in terms of distance error.

show abstract

Section: Growing Neural Gas Networkmentioning

confidence: 99%

Section: Growing Neural Gas Networkmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Performance analysis of self-organising neural networks tracking algorithms for intake monitoring using kinect

Gasparrini¹,

Cippitelli²,

Gambi³

et al. 2015

IET International Conference on Technologies for Active and Assisted Living (TechAAL)

View full text Add to dashboard Cite

show abstract

“…• Migniot et al [35] address the problem of 3D tracking of human gesture for buying behavior estimation. The top view of the customers, which has not received much attention in human tracking, is exploited in this particular context.…”

Section: In the Context Of Real-time Tracking And Navigation Systemsmentioning

confidence: 99%

Real-time motion estimation for image and video processing applications

Botella

Garcı́a

2014

J Real-Time Image Proc

View full text Add to dashboard Cite

This contribution focuses on different topics that are covered by the special issue titled ''Real-Time Motion Estimation for image and video processing applications'' and which incorporate GPUS, FPGAs, VLSI systems, DSPs, and Multicores, among other platforms. The guest editors have solicited original contributions, which address a wide range of theoretical and practical issues related to high-performance motion estimation image processing including, but not limited to: real-time matching motion estimation systems, real-time energy-based motion estimation systems, gradient-based motion estimation systems, optical flow estimation systems, color motion estimation systems, multi-scale motion estimation systems, optical flow and motion estimation systems, analysis or comparison of specialized architectures for motion estimation systems and realworld applications.

show abstract

“…However, for recovering the 3D state using many views one may attack the problem as a collection of 2D single views (evaluating the pose individually for each view) and integrating those views geometrically or, more accurately, voxelizing the space and evaluating the 3D positioning and tracking [22]. Another approach considers a depth acquisition sensor (such as Kinect-like devices) as in [21]. Our hybrid approach, also taken in [22], uses a 3D particle filter and fully calibrated cameras.…”

Section: Introductionmentioning

confidence: 99%

Performance evaluation of a 3D multi-view-based particle filter for visual object tracking using GPUs and multicore CPUs

Concha

Cabido

Pantrigo

et al. 2014

J Real-Time Image Proc

View full text Add to dashboard Cite

This paper presents a deep and extensive performance analysis of the particle filter (PF) algorithm for a very compute intensive 3D multi-view visual tracking problem. We compare different implementations and parameter settings of the PF algorithm in a CPU platform taking advantage of the multithreading capabilities of the modern processors and a graphics processing unit (GPU) platform using NVIDIA CUDA computing environment as developing framework. We extend our experimental study to each individual stage of the PF algorithm, and evaluate the quality versus performance trade-off among different ways to design these stages. We have observed that the GPU platform performs better than the multithreaded CPU platform when handling a large number of particles, but we also demonstrate that hybrid CPU/GPU implementations can run almost as fast as only GPU solutions.

show abstract

Hybrid 3D–2D human tracking in a top view

Cited by 23 publications

References 43 publications

Performance analysis of self-organising neural networks tracking algorithms for intake monitoring using kinect

Performance analysis of self-organising neural networks tracking algorithms for intake monitoring using kinect

Real-time motion estimation for image and video processing applications

Performance evaluation of a 3D multi-view-based particle filter for visual object tracking using GPUs and multicore CPUs

Contact Info

Product

Resources

About