Emperor Penguin optimized event recognition and summarization for cricket highlight generation

Shingrakhia, Hansa; Patel, Hetal

doi:10.1007/s00530-020-00684-3

Cited by 14 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using 3D CNNs (3 Dimensional Convolutional Neural Networks) and LSTM (Long Short-Term Memory) to extract features from a soccer video for summarization [5] & [6]. Shingrakhia et al [7,8] have mentioned in their research work about video summarization techniques using novel deep learning algorithms such as SGRNN-AM (Stacked Gated Recurrent Neural Network with Attention Module) and HDNN-EPO (Hybrid Deep Neural Network with Emperor Penguin Optimization) for improved performance on cricket videos with 96% accuracy and precision. J. Yu et al [9] have designed an architecture that detects events and proposes temporal boundaries to the events to form what they call stories i.e.…”

Section: Literaturementioning

confidence: 99%

Video Summarization for Multiple Sports Using Deep Learning

Guntuboina¹,

Porwal²,

Jain³

et al. 2021

ELCVIA

Self Cite

View full text Add to dashboard Cite

This paper proposes a computationally inexpensive method for automatic key-event extraction and subsequent summarization of sports videos using scoreboard detection. A database consisting of 1300 images was used to train a supervised-learning based object detection algorithm, YOLO (You Only Look Once). Then, for each frame of the video, once the scoreboard was detected using YOLO, the scoreboard was cropped out of the image. After this, image processing techniques were applied on the cropped scoreboard to reduce noise and false positives. Finally, the processed image was passed through an OCR (Optical Character Recognizer) to get the score. A rule-based algorithm was run on the output of the OCR to generate the timestamps of key-events based on the game. The proposed method is best suited for people who want to analyse the games and want precise timestamps of the occurrence of important events. The performance of the proposed design was tested on videos of Bundesliga, English Premier League, ICC WC 2019, IPL 2019, and Pro Kabaddi League. An average F1 Score of 0.979 was achieved during the simulations. The algorithm is trained on five different classes of three separate games (Soccer, Cricket, Kabaddi). The design is implemented using python 3.7.

show abstract

Section: Literaturementioning

confidence: 99%

Video Summarization for Multiple Sports Using Deep Learning

Guntuboina¹,

Porwal²,

Jain³

et al. 2021

ELCVIA

Self Cite

View full text Add to dashboard Cite

show abstract

“…(3) Nonmaximum suppression of gradient amplitude: the above process only gets the global gradient is not enough to determine the edge; this step that is to refine the image of the gradient amplitude of the extreme value region, retain the local gradient amplitude of the largest point, to suppress the nonmaximum value, to achieve the effect of refining the edge. First, based on the neighborhood of G as the center of the direction, the neighborhood is interpolated in the G gradient direction [15], and then, the center point G is compared with the interpolation size of the two adjacent gradient magnitudes in its gradient direction, and if the magnitude of the center point of the neighborhood is greater than or equal to the interpolation result in its gradient direction, it can be regarded as the initial edge point, and if it is less than this value, it is regarded as a nonedge point, and its value is regarded as zero, and after passing this step, the image can be obtained. And a suitable method for sports event optimization can be obtained.…”

mentioning

confidence: 99%

Optimization of Sports Event Management System Based on Wireless Sensor Network

et al. 2021

View full text Add to dashboard Cite

In recent years, wireless sensor network technology has developed rapidly and its role in managing systems for sports events has been widely used. Wireless sensor networks not only have low wiring cost, high monitoring accuracy, and good fault tolerance but also can be monitored remotely and have outstanding advantages in fault diagnosis and safety monitoring. In this paper, firstly, the wireless sensor network hierarchical routing protocol is studied and its network model and workflow are analyzed; according to the energy consumption of the wireless sensor network, the selection method of the optimal number of cluster heads is proposed to analyze the advantages and disadvantages existing in the protocol. Secondly, the improvement of the routing protocol is proposed to address the problems of uneven distribution of cluster heads and cluster head election without considering the residual energy of nodes in the protocol. When dividing clusters, the number of neighboring nodes is considered so that cluster heads are distributed more evenly in the network; when electing cluster heads, the residual energy of nodes in the cluster is considered to balance the whole network load, and when electing cluster heads, the residual energy of nodes in the cluster is considered to balance the whole network load. Finally, simulation experiments are conducted in this paper using simulation software, and the simulation results show that the data fusion algorithm is more effective than the protocol in reducing the average energy consumption of nodes and extending the network lifetime; these features make wireless sensors more beneficial for better management of sporting events as well as better optimization.

show abstract