Emerging technologies are rapidly transforming the way we interact with media and how we leverage it for business advantage. Although companies store media for many legal and business purposes, many media archives are underutilized due to the labor-intensive process of metadata tagging, indexing, and updating older files. The limitations of typical keyword searches make it challenging to find relevant footage that can be leveraged to derive business value promptly.
Today, rapid media retrieval and editing are becoming more accessible thanks to the advent of Artificial Intelligence (AI). By leveraging machine learning (ML) and image recognition, AI can be trained for various applications, creating robust metadata associated with the archives and acting as an accelerator for specific use cases, thereby enhancing the time-to-value. Coupling this with large language models (LLMs) can provide a robust solution, enabling users to search through extensive media archives easily and quickly. To showcase the evolving capabilities of AI, FPT Software developed its SportsCut Pro concept.
FPT’s SportsCut Pro concept leverages publicly available MLB footage and is trained to recognize players, teams, and actions within a game, and ties those actions to events within each game. The LLM enables natural language searches that can be refined until the user gets the clips that they are looking for. Enhanced metadata – such as the ability to automatically identify and tag key plays in a sporting event or identify sentiment in movie scenes – makes it easier to find and use archival assets in new productions, accelerating time to market and enabling rapid response to content demands.
For example, you can search for “Shohei Ohtani at bats in 2024”, then refine that to “Only against the NY Yankees”, and then further refine it to “Show me only home runs” - to return a set of video clips highlighting Ohtani’s home runs against the Yankees. Enhanced metadata – such as the ability to automatically identify and tag key plays in a sporting event or identify sentiment in movie scenes – makes it easier to find and use archival assets in new productions, accelerating time to market and enabling rapid response to content demands.
A natural first-adopter of this technology is the entertainment, sports, and media industries. Media companies can repurpose archived footage, TV episodes, documentaries, and raw material for new programming, documentaries, or digital distribution, increasing audience reach and generating new revenue streams. Archival content can be licensed to third parties or monetized on streaming platforms, capitalizing on trends such as the growing demand for documentaries or nostalgic content. A natural user group is sports editors who can now quickly find and insert media clips from legacy athletic performances to build compelling digital stories.
This concept illustrates the potential of AI to enable organizations to unlock new value from decades of stored media archives. “We developed SportsCut Pro to demonstrate the art of the possible,” says Ira Dworkin, Managing Director, CME for FPT Software. “This cloud-agnostic solution, currently residing on the Akamai Cloud Platform, can be trained to identify various actions and people in videos, providing solutions for a wide variety of use cases across various industries.” Imagine leveraging AI to rapidly surface videos that may be locked in over a hundred years of archived sports or movie footage.
While the SportsCut Pro concept is initially focused on helping the entertainment and sports industries derive value from video archives, the technology holds promise for other sectors.
Journalism and News Production.
News organizations, particularly those in broadcast news, can utilize this technology to search and retrieve relevant video footage for current stories easily, analyze historical trends, and create richer, more compelling journalism. Archival footage is essential for producing retrospectives, anniversary coverage, and investigative pieces that require historical context.
Enterprise and Training Applications
Businesses can employ a SportsCut Pro-like solution to extend the ‘art of the possible’ to enterprise and training environments. The solution can partner with existing enterprise systems to accelerate content discovery, improve knowledge transfer, reduce manual effort, support compliance, and ultimately maximize the value of digital media. It can seamlessly integrate with enterprise content management, customer relationship management, and learning systems to enhance workflows. Furthermore, its use of improved analytics can lead to continuous learning, filling content gaps, and training courses, as well as optimizing knowledge sharing.
Brand Storytelling and Marketing
Increasingly, business is conducted online, and the digital production of presentations, product launches, and events is becoming the standard. Archived media can be leveraged to craft compelling brand stories, highlight a company's history, and reinforce its values, thereby building reputation and fostering customer loyalty. Marketing and advertising teams benefit from quick access to historical campaign assets, enabling consistent branding and efficient reuse of content across campaigns.
Down the Road
SportsCut Pro was developed to highlight what’s possible. When coupled with FPT’s extensive experience in AI, it’s not hard to imagine the concept being applied to the search and retrieval of media beyond video. This will significantly enhance operational efficiency for many use cases. Additionally, each application is trained on data specific to each customer’s use case, with the resulting product wholly owned by the customer.
The Sports Cut Pro concept is just the beginning of many ways customers can discover the art of the possible to enhance the optimization of archived media value. Today, we can apply it to video retrieval, but as we partner with customers, we’ll discover new ways to leverage and preserve media for greater value.
You Might Like: Understand more about ML Infrastructure and its Best Practices
