<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>In this paper, the classical online multi-target tracking algorithms SORT and DeepSORT are improved. Firstly, we reduce the frequency of ID-Switch occurrence by integrating representational information, on which we complete the improvement of SORT to obtain the DeepSORT algorithm. Then, an MF-SORT multi-target tracking algorithm is proposed by using the squared Martens distance instead of the cosine distance of the high-dimensional representational features in DeepSORT to track the similarity between the frame and the detection frame. Subsequently, the algorithm is applied to analyze soccer training videos, and in this way, a soccer teaching and training method is proposed. The results of the teaching experiment show that the average score of the players who receive this method is 2.7 points ahead of the players who receive the traditional method on average. The post-test score of the experimental group is 2.6 points higher than the pre-test score. The players who received the training method were 10.2% more satisfied with their training than the players who did not receive it. It shows that the video analysis method for ball training has good results both in soccer training performance and satisfaction with training.</p>
</abstract>

In this paper, the classical online multi-target tracking algorithms SORT and DeepSORT are improved. Firstly, we reduce the frequency of ID-Switch occurrence by integrating representational information, on which we complete the improvement of SORT to obtain the DeepSORT algorithm. Then, an MF-SORT multi-target tracking algorithm is proposed by using the squared Martens distance instead of the cosine distance of the high-dimensional representational features in DeepSORT to track the similarity between the frame and the detection frame. Subsequently, the algorithm is applied to analyze soccer training videos, and in this way, a soccer teaching and training method is proposed. The results of the teaching experiment show that the average score of the players who receive this method is 2.7 points ahead of the players who receive the traditional method on average. The post-test score of the experimental group is 2.6 points higher than the pre-test score. The players who received the training method were 10.2% more satisfied with their training than the players who did not receive it. It shows that the video analysis method for ball training has good results both in soccer training performance and satisfaction with training.

Optimization method of soccer sports teaching and training assisted by intelligent technology

Jiyang College of ZheJiang A&F University

Applied Mathematics and Nonlinear Sciences

This work is licensed under the Creative Commons Attribution 4.0 International License.

In this paper, the classical online multi-target tracking algorithms SORT and DeepSORT are improved. Firstly, we reduce the frequency of ID-Switch occurrence...

Optimization method of soccer sports teaching and training assisted by intelligent technology

Pubblicato online: 05 lug 2024

Pagine: -

Ricevuto: 01 mar 2024

Accettato: 20 mag 2024

DOI: https://doi.org/10.2478/amns-2024-1696

Parole chiave
SORT, DeepSORT, Squared Marginal Distance, MF-SORT, Soccer training

© 2024 Pengtao Cui et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Optimization method of soccer sports teaching and training assisted by intelligent technology

Pubblicato online: 05 lug 2024

Pagine: -

Ricevuto: 01 mar 2024

Accettato: 20 mag 2024

DOI: https://doi.org/10.2478/amns-2024-1696

Parole chiaveSORT, DeepSORT, Squared Marginal Distance, MF-SORT, Soccer training

© 2024 Pengtao Cui et al., published by Sciendo

This work is licensed under the Creative Commons Attribution 4.0 International License.

Parole chiave
SORT, DeepSORT, Squared Marginal Distance, MF-SORT, Soccer training