- 期刊
初探科技輔助介入棒球主審好壞球判決之影響
A preliminary study on the impact of technology assistance on the judgment of baseball referee's strike zone
摘要
在體育運動上，結合科技輔助的案例及構想漸漸成為話題，人工智慧領域的發展推動著社會的進步，而其中圖像識別近年受到各界關注。圖像識別中的目標檢測技術讓人們實現多項高科技應用，而YOLO模型提高了目標檢測的效率，使得人們能建構出系統以完成即時目標檢測。本文將談論關於機器學習與深度學習之差異，簡介關於YOLO模型的深度學習方法，將YOLO模型應用於好球帶的判斷，並對於人工智慧應用於體育運動進行探討。在瞭解人工智慧中的深度學習之物件偵測系統YOLO之後，分析其套用在棒球裁判訓練或比賽實務中的方法，依科技理論，此研究方向具高可行性，YOLO系統的運算能力及速度，具有相當大的優勢，能實際的執行即時運算。建議未來研究方向，將撰寫程式碼，經機器學習後，建立YOLO模型，實際於棒球比賽中執行，對其準確率進行分析，並加以改善，以實現真正的即時科技輔助判決。
並列摘要
In sports, cases and ideas combined with science and technology have gradually become a topic. The development of artificial intelligence has promoted the progress of society, and image recognition has attracted attention from all walks of life in recent years. The target detection technology in image recognition allows people to achieve a number of high-tech applications, and the YOLO model improves the efficiency of target detection, enabling people to construct a system to complete real-time target detection. This article will talk about the difference between machine learning and deep learning, introduce the deep learning method of the YOLO model, apply the YOLO model to the judgment of the strike zone, and discuss the application of artificial intelligence to sports. After understanding the object detection system YOLO, which is a deep learning in artificial intelligence, analyze its application in baseball referee training or game practice. According to scientific and technological theory, this research direction is highly feasible. The computing power and speed of the YOLO system have considerable advantages and can actually perform real-time computing. It is recommended that the future research direction is to write code, build a YOLO model after machine learning, and implement it in a baseball game. It is accuracy and improve to achieve real real-time technology-assisted judgments.