Hockey Puck Tracking Challenge

🛠 Contribute : Found a typo? Something confusing? Have a suggestion? Please reach out

🕵️ Introduction

The goal of this challenge is to process an image or video of a hockey game and extract the x and y coordinates of the hockey puck.

⏱ Timeline

• Start Time: 4:00 PM  (UTC) 25th November 2020
• End Time: 4:00 PM (UTC) 15th February 2021

🏆 Prizes

There will be 3 cash prizes;

1. $2,500 USD
2. $1,000 USD
3. $500 USD

Eligibility for cash prizes requires that the final solution performs relative to the score acheived on a second dataset of frames. 

💾 Dataset

The dataset consists of videos and images.  The videos vary in view point, camera type and lighting.  The dataset folder is structured in the following way:

• images
  • <videoId>
    • <frame>.jps
• videos
  • <videoId>.ext

🚀 Submission

• Prepare a csv containing the following headers
  • videoId ~ This can be extracted from the dataset path dataset/{id}.ext
  • frameId ~ This is the frame id that coordinates are extracted from
  • puck ~ This is the coordinates in "[x,y]" format
  • time (ms) ~ The time in milliseconds it took to run the code that processed the frame

👩‍💻 Getting Started

Under the resource tab there is a skeleton implementation. It provides a python boilerplate code that calculates time of execution and prepares the information needed both for image processing and/or video processing.

🖊 Evaluation Criteria

Primary Scoring

Primary scoring is calculated using over 1,000 ground truth results and calculating the distance provided from the ground truth.  The closer to the ground truth the larger the score for that frame.

Secondary Scoring

The secondary scoring is the sum of the time (ms) column, faster solutions will win tie breakers

📱 Contact

• Jason Brumwell