to identify the speed of the pitched ball that is displayed…
to identify the speed of the pitched ball that is displayed in broadcast videos of MLB using machine learning. please Attach a proposal according to above question addressing the following: Objectives Books and materials used Brief description of project Deliverables Evaluation criteria
Answer
Proposal: Identifying the Speed of Pitched Balls in MLB Broadcast Videos Using Machine Learning
Objectives:
The objective of this study is to develop a machine learning model that can accurately identify the speed of pitched balls in broadcast videos of Major League Baseball (MLB) games. Specifically, the model will be trained to analyze video footage and extract the relevant features necessary to estimate the speed of the ball at the moment it crosses the catcher’s glove. The ultimate goal is to provide a reliable and automated method for determining pitch speed from broadcast videos in real-time.
Books and Materials Used:
In order to accomplish the objectives of this study, the following books and materials will be utilized:
1. “Machine Learning: A Probabilistic Perspective” by Kevin P. Murphy – This book provides a comprehensive introduction to machine learning algorithms and techniques, covering both the theoretical foundations and practical implementation aspects.
2. “Computer Vision: Algorithms and Applications” by Richard Szeliski – This book serves as a comprehensive guide to computer vision techniques, including topics such as image recognition, object detection, and motion analysis, which are relevant to our project.
3. A collection of MLB broadcast videos – These videos will be used as the primary dataset to train and evaluate the machine learning model. The videos will be obtained from public sources or with permission from MLB.
Brief Description of Project:
The project aims to leverage machine learning algorithms and computer vision techniques to accurately estimate the speed of pitched baseballs in MLB broadcast videos. The proposed workflow involves the following steps:
1. Data Collection: A collection of varied MLB broadcast videos will be obtained, covering a wide range of pitching scenarios and camera angles. These videos will serve as the dataset for training and evaluating the machine learning model.
2. Data Preprocessing: The videos will be preprocessed to extract frames containing the pitch trajectory. Techniques such as frame interpolation and image enhancement may be applied to enhance the quality and resolution of the frames.
3. Feature Extraction: Using computer vision techniques, relevant features will be extracted from each frame, including the trajectory of the ball, the positions of the pitcher and the catcher, and other contextual information.
4. Training and Validation: The extracted features will be used to train a machine learning model, such as a convolutional neural network or a recurrent neural network, to predict the speed of the pitch. The trained model will be validated using a separate set of videos.
Deliverables:
The deliverables of this project will include:
1. A machine learning model trained to accurately estimate the speed of pitched balls in MLB broadcast videos.
2. A user-friendly software application that utilizes the trained model to analyze and provide real-time pitch speed measurements from broadcast videos.
3. A comprehensive report documenting the methodology, experimental results, and implementation details of the project.
Evaluation Criteria:
The performance of the machine learning model will be evaluated based on its accuracy in estimating the speed of pitched balls. The evaluation will be conducted by comparing the model’s predictions with the ground truth pitch speed data available from official sources, such as radar guns or pitch tracking systems. Metrics such as mean absolute error and mean square error will be used to quantify the model’s accuracy. Additionally, the real-time processing capabilities of the software application will also be evaluated to ensure its practicality in live broadcast scenarios.