Real-time event classification in field sport videos

Kapela, Rafał; Świetlicka, Aleksandra; Rybarczyk, Agnieszka; Kolanowski, Krzysztof; O’Connor, Noel E.

doi:10.1016/j.image.2015.04.005

Cited by 16 publications

(14 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A. O'Reilly, Whelan, Ward, Delahunt, & Caulfield, 2017b;Ó Conaire et al, 2010;Reily et al, 2017;Shah et al, 2007;Zhu et al, 2006). Other algorithms included kNN (n = 3) (Díaz-Pereira et al, 2014;Montoliu et al, 2015;Ó Conaire et al, 2010), decision tree (DT) (n = 2) (Kapela et al, 2015;Liao et al, 2003), RF (n = 2) (Kasiri-Bidhendi et al, 2015;Kasiri et al, 2017), and Multilayer Perceptron (MLP) (n = 2) (Kapela et al, 2015;Montoliu et al, 2015). Deep learning was investigated in seven studies (Bertasius et al, 2017;Ibrahim, Muralidharan, Deng, Vahdat, & Mori, 2016;Karpathy et al, 2014a;Nibali et al, 2017;Ramanathan et al, 2015;Tora, Chen, & Little, 2017;Victor et al, 2017) of which used CNNs or LSTM RNNs as the core model structure.…”

Section: Vision Recognition Model Development Methodsmentioning

confidence: 99%

“…One study (Karpathy et al, 2014a) used the publicly available Sports-1M, as previously described. Vision-based studies also reported datasets in total time, 10.3 hours (Bertasius, Park, Yu, & Shi, 2017), 3 hours (Montoliu, Martín-Félez, Torres-Sospedra, & Martínez-Usó, 2015), 1, 500 minutes (Shah et al, 2007), and 50 hours (Kapela et al, 2015), and by frame numbers, 6, 035 frames (Zhu, Xu, Gao, & Huang, 2006) and 10, 115 frames (Reily et al, 2017).…”

Section: Experimental Designmentioning

confidence: 99%

“…Well-established filters for processing motion signal data include the Kalman filter 133 Titterton & Weston, 2009;D. Wagner, Kalischewski, Velten, & Kummert, 2017) and a Fourier 134 transform filter such as a fast Fourier transform 135 (Kapela, Świetlicka, Rybarczyk, Kolanowski, & O'Connor, 2015;Preece, Goulermas, Kenney, & 136 Howard, 2009). Near real-time processing benefits from reducing memory requirements, 137 computational demands, and essential bandwidth during whole model implementation.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Machine and deep learning for sport-specific movement recognition: a systematic review of model development and performance

Cust

Sweeting

Ball

et al. 2018

Journal of Sports Sciences

211

148

View full text Add to dashboard Cite

Objective assessment of an athlete's performance is of importance in elite sports to facilitate detailed analysis. The implementation of automated detection and recognition of sport-specific movements overcomes the limitations associated with manual performance analysis methods. The object of this study was to systematically review the literature on machine and deep learning for sport-specific movement recognition using inertial measurement unit (IMU) and, or computer vision data inputs. A search of multiple databases was undertaken. Included studies must have investigated a sportspecific movement and analysed via machine or deep learning methods for model development. A total of 52 studies met the inclusion and exclusion criteria. Data preprocessing, processing, model development and evaluation methods varied across the studies. Model development for movement recognition were predominantly undertaken using supervised classification approaches. A kernel form of the Support Vector Machine algorithm was used in 53% of IMU and 50% of vision-based studies. Twelve studies used a deep learning method as a form of Convolutional Neural Network algorithm and one study also adopted a Long Short Term Memory architecture in their model. The adaptation of experimental setup , data pre-processing, and model development methods are best considered in relation to the characteristics of the targeted sports movement(s).

show abstract

Section: Vision Recognition Model Development Methodsmentioning

confidence: 99%

Section: Experimental Designmentioning

confidence: 99%

mentioning

confidence: 99%

See 1 more Smart Citation

Machine and deep learning for sport-specific movement recognition: a systematic review of model development and performance

Cust

Sweeting

Ball

et al. 2018

Journal of Sports Sciences

211

148

View full text Add to dashboard Cite

show abstract

“…In sports broadcasts the cameras and the camera viewpoint change frequently. For example, in a short period a camera might show the broad view of the field, followed by a close-up of the player and then move to crowd reaction [7]. Therefore in any sports analysis system, first the video is broken down into different shots, and then the shots are analysed to extract the structural information for that particular sport.…”

Section: Reviewmentioning

confidence: 99%

Scene structure analysis for sprint sports

Hedayati

Cree

Scott

2016

2016 International Conference on Image and Vision Computing New Zealand (IVCNZ)

View full text Add to dashboard Cite

This work proposes a robust model to analyse the structure of horse races based on 2D velocity vector information. This model is capable of detecting scene breaks, classifying the view of the contenders and extracting the trajectory of the contenders throughout the race. The performance of the system is tested over six video clips from two different broadcast sources. The performance analysis shows the model achieves a high accuracy of view classification with the lowest value of 83%, all in real time.

show abstract

“…These treatments involve a large variety of complex techniques and necessitate high processing power. Real-time video content analysis and its integration in embedded systems are also becoming a focus of research [2,[4][5][6]. Video content analysis techniques are used in constrained applications [2,3] like real-time content delivery, interactive TV, * Correspondence: abdessalem.benabdelali@enim.rnu.tn surveillance systems, security systems, and personal video recorders.…”

Section: Introductionmentioning

confidence: 99%

FPGA-based SOC for hardware implementation of a local histogram-based video shot detector

Abdelali¹,

Krifa²,

Mtibaa³

2017

Turk J Elec Eng & Comp Sci

View full text Add to dashboard Cite

Abstract:In this paper, we present a video application example and its implementation in a reconfigurable system-onchip (SOC) platform. The proposed platform employs the benefits of field programmable gate array (FPGA) technology.A prototype based on a Xilinx Virtex-5 FPGA is developed. The application includes a video shot boundary detection module based on the local histogram (LH) technique. Diverse hardware module versions corresponding to different quantization levels and architectural solutions for an LH-based shot detection system are presented. The developed modules have different hardware resource occupations and can be used in a dynamic way to allow flexible management of the target hardware system. They also show high execution time efficiency and can reach an important bandwidth that can support the most recent high-resolution video formats with a high rate of frames per second. A complete SOC-based demonstration for video summarization based on the LH is also developed. It includes the LH extraction, shot boundary detection, and key frame visualization.

show abstract

Real-time event classification in field sport videos

Cited by 16 publications

References 39 publications

Machine and deep learning for sport-specific movement recognition: a systematic review of model development and performance

Machine and deep learning for sport-specific movement recognition: a systematic review of model development and performance

Scene structure analysis for sprint sports

FPGA-based SOC for hardware implementation of a local histogram-based video shot detector

Contact Info

Product

Resources

About