We are making a desktop application which will generate cricket highlights using the video of cricket match on the actions of umpire. It’s a better and easy way of making highlights automatically. As there are many video highlights generator on internet but we can’t find a proper application for cricket matches. Motivated to extract actions of umpire in cricket videos in the most efficient and effective manner, we want to obtain the main clips of match. And also the clips we will provide against the full video will be helpful for highlights of match.
Objectives
The idea was to automate the current system by detecting the actions on cricket match video.
Specifically, the objectives of this project were:
• To design and implement an approach for action detection
• To analyse cricket match.
• To generate video highlights.
• To use collection of images as training set.
• To design Algorithm and implementation of object detection and classification.
• To design Algorithm for Umpire detection.
• To design Algorithm for Umpire’s Action detection and classification
Socio-Economic Benefit
It will provide the facility to generate clips of some specific scenarios in cricket match, performing all the basic learning methodology .
The system can be used in the sports channel and news channel who can use our system to gather clips of specific actions and save a lot of time and reducing number of employees.
Methodologies
The Agile methodology was used in the development of the Auto Cricket Highlight Generator project. This methodology emphasizes collaboration, flexibility, and the delivery of a working product in short iterations. The project was divided into several sprints.
The Agile methodology was chosen because it allows for greater flexibility and adaptability in the development process. The methodology allows for the group members to respond quickly to changes in the requirements or priorities of the project.
Overall, the Agile methodology was a successful approach for the development of the NFT Horizon project, allowing for efficient and effective development with a focus on delivering a quality product that meets the needs of the users.
Outcome
Outcome of project is achieved that is highlight generation .Apply the algorithms to a real problem, optimize the models learned and report on the expected accuracy that can be achieved by applying the models. We want to introduce and create a system that could automate the video highlights by recognizing actions in cricket match, which is cheap and fully automated. This was the main reason for developing the system. This system uses object classification with the accuracy of 74.0% and also key feature of this system is Faster R-CNN to detect actions in cricket match.
