An improvement on an MCMC-based video tracking algorithm

Shalnov, Evgeny; Konushin, Vadim; Konushin, Anton

doi:10.1134/s1054661815030220

Cited by 6 publications

(4 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We will research applications of probabilistic convolutional networks to human pose estimation in video. human pose estimation in video via MCMC sampling by Shalnov and Konushin [25] will be used as a baseline for future works.…”

Section: Discussionmentioning

confidence: 99%

Deep probabilistic human pose estimation

Petrov

Shakhuro

Konushin

2018

IET Computer Vision

View full text Add to dashboard Cite

The authors consider the problem of human pose estimation using probabilistic convolutional neural networks. They explore ways to improve human pose estimation accuracy on standard pose estimation benchmarks MPII human pose and Leeds Sports Pose (LSP) datasets using frameworks for probabilistic deep learning. Such frameworks transform deterministic neural network into a probabilistic one and allow sampling of independent and equiprobable hypotheses (different outputs) for a given input. Overlapping body parts and body joints hidden under clothes or other obstacles make the problem of human pose estimation ambiguous. In this context to get accurate estimation of joints' position they use uncertainty in network's predictions, which is represented by variance of hypotheses, provided by a probabilistic convolutional neural network, and confidence is characterised by mean of them. Their work is based on current CNN cascades for pose estimation. They propose and evaluate three probabilistic convolutional neural networks built on top of deterministic ones with two probabilistic deep learning frameworks-DISCO networks and Bayesian SegNet. The authors evaluate their models on standard pose estimation benchmarks and show that proposed probabilistic models outperform base deterministic ones.

show abstract

Section: Discussionmentioning

confidence: 99%

Deep probabilistic human pose estimation

Petrov

Shakhuro

Konushin

2018

IET Computer Vision

View full text Add to dashboard Cite

show abstract

“…There are a lot of ways to detect the desired object on the frame. Three most popular methods are: detection of body [2,3,4], detection of head [5,6] and using of key points. The first solution is popular because there are many datasets and ready-made solutions.…”

Section: Related Workmentioning

confidence: 99%

A Distributed Tracking Algorithm for Counting People in Video by Head Detection

Kuplyakov

Geraskin

Mamedov

et al. 2020

Proceedings of the 30th International Conference on Computer Graphics and Machine Vision (GraphiCon 2020). Part 2

View full text Add to dashboard Cite

We consider the problem of people counting in video surveillance. This is one of the most popular tasks in video analysis because this data can be used for predictive analytics and improvement of customer services, traffic control, etc. Our method is based on the object tracking in video with low framerate. We use the algorithm from [1] as a baseline and propose several modifications that improve the quality of people counting. One of the main modifications is to use a head detector instead of a body detector in the tracking pipeline. Head tracking is proved to be more robust and accurate as the heads are less susceptible to occlusions. To find the intersection of a person with a signal line, we either raise the signal lines to the level of the heads or perform a regression of bodies based on the available head detections. Our experimental evaluation has demonstrated that the modified algorithm surpasses the original in both ac- curacy and computational efficiency, showing a lower counting error on a lower detection frequency.

show abstract

“…The proposed approach uses evidence of the pose from the other frames for the result gaining in the current frame. We use the tracking approach (Shalnov and Konushin, 2013) to initially estimate trajectory of the person. A trajectory means an approximate location of the person in each frame of the input video sequence.…”

Section: Task Definitionmentioning

confidence: 99%

“…Our algorithm of human pose estimation in a video isn't restricted to a choice from the set of hypotheses as well. The specific choice of the inference method was inspired by successful usage of the sampling technique for the tracking task (Shalnov and Konushin, 2013). In addition, it makes inference in complicated mathematical models possible.…”

Section: Related Workmentioning

confidence: 99%

Human Pose Estimation in Video via MCMC Sampling

Shalnov¹,

Konushin²

2015

Proceedings of the 5th International Workshop on Image Mining. Theory and Applications

View full text Add to dashboard Cite

We describe a method for the human pose estimation in a video sequence. We propose a new mathematical model of a human pose in a video sequence, which incorporates motion and pose parameters. We show that the model of (Park and Ramanan, 2011) is a particular case of our model. We introduce a framework to infer an approximation of the optimal value in the proposed model. We use an exact algorithm of motion parameters estimation to reduce complexity of inference. Our approach outperforms results of (Park and Ramanan, 2011) in the most complicated video sequences.

show abstract

An improvement on an MCMC-based video tracking algorithm

Cited by 6 publications

References 3 publications

Deep probabilistic human pose estimation

Deep probabilistic human pose estimation

A Distributed Tracking Algorithm for Counting People in Video by Head Detection

Human Pose Estimation in Video via MCMC Sampling

Contact Info

Product

Resources

About