Skip to content

Opening book details…

Can I read 2019 14th International Joint Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP) - Automatic Football Match Event Detection from the Scoreboard using a Single-Shot MultiBox Detector on EtoBox?

2019 14th International Joint Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP) - Automatic Football Match Event Detection from the Scoreboard using a Single-Shot MultiBox Detector by Somwong, Rungroj; Hnoohom, Narit is a scholarly article available to read on EtoBox.

What is 2019 14th International Joint Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP) - Automatic Football Match Event Detection from the Scoreboard using a Single-Shot MultiBox Detector about?

During a football match, the information is manually collected by humans. However, the correctness of the football match data is difficult to check because of the game's speed, and thus, human errors can occur. This paper presents an automatic football match event detection from the scoreboard using a deep learning algorithm. The proposed method can reduce human error and performs the detection faster. In this study, the detection was trained with 30,000 data of Goals, Substitutions and Cards scoreboard from 68 matches of English Premier League 2017-2018 broadcast videos. The detection was tested with 80 sub-testing videos. These videos were prepared from 20 full matches broadcast videos, which consisted of 12 full matches from the year 2017-2018 and 8 full matches from the year 2018-2019. The proposed method contains three main steps: data gathering and augmentation, object detection for scoreboard visualization forms, and the event classification. The scoreboard detection is performed with a Single-Shot MultiBox Detector. The event classification employs the majority vote and time frame technique. The experimental results show an accuracy rate of 1.00 with the expected event scor

Author
Somwong, Rungroj; Hnoohom, Narit
Publisher
IEEE
Published
2019
Language
EN

More by Somwong, Rungroj; Hnoohom, Narit

Browse all works by Somwong, Rungroj; Hnoohom, Narit