Skip to content

Market Overview

Sports are one of the most prominent and profitable forms of entertainment in the United States; 102 million viewers tuned in to watch the Super Bowl this year, and 25.5 million watched the CFP National Championship. Additionally, everyone wants to represent their favorite teams; the sports apparel market is booming, valued at $167.7 billion in 2018, and projected to grow to $248.1 billion by 2026. However, when broadcasting games, brand owners, networks, and fans all face a similar problem with brand color consistency across platforms. Current color-correction solutions adjust the entire frame, which can negatively impact the color of everything else including skin tones and field conditions. To solve this problem, Clemson Researchers developed ColorNet, a novel software that uses artificial intelligence to identify colors important to a brand, such as team colors on a jersey, and correct them to the brand-correct color, in real time.


Color Correction, Artificial Intelligence, Neural Networks

Technical Summary:

Optimizing the display of brand colors on large-format video displays during live sporting events is a complex process that is usually handled manually by specially trained professionals. Color visualization on video displays is impacted by a variety of factors, including the video source, the screen display, and the ambient light at the location of the event. This is usually resolved by manually adjusting the amount of red, green, and blue (RGB) in the video, however, on big game days there can be up to 50 cameras making it difficult to maintain visual consistency and brand correct colors in real time across all feeds. This specialized AI solution targets only the specified brand colors, and adjusts those areas of each video frame across all camera inputs, creating a much more cohesive display.


  • The process is automated, circumventing human error when color correcting for brand colors
  • The process functions in real time, minimizing the negative impact of lighting or weather changes on the display of brand colors
  • The neural network is trained to identify brand colors off a small data set, conferring the ability to target specific coors and adjust pixel-by-pixel without shifting any other surrounding colors

Download Printable PDF

Technology Overview

State of Development

Production level prototype

Patent Type



Information and Communication

Serial Number


CURF Reference No.



Dr. Erica Walker, Dr. Hudson Smith

For More Info, Contact:

Interested in this technology?
Contact curf@clemson.edu
Please put technology ID in subject line of email.


Get Started
Ready to Get Started?

Contact our team at CURF

Contact CURF