System can cut long-form videos into short clips

Pictory logo

Pictory has gained a US patent for its AI-powered video synopsis technology, after a three-year application process. 

The model utilizes a novel algorithm and custom large language models to generate video summaries from transcriptions. When a video is uploaded, the spoken words are transcribed into text. Then, these sentences are grouped into topics based on their similarity. The topics are ranked and clustered, creating a cohesive summary. By using advanced algorithms, Pictory automatically extracts key points from video transcriptions.

This can be used to turn long-form videos into short clips, potentially for use as highlights or previews. Companies such as WSC Sports and Magnifi offer similar services for the media and entertainment market.

Vishal Chalana, chief technology officer at Pictory, said: “We are immensely proud to announce that the U.S. Patent & Trademark Office has officially recognized our team’s innovation and dedication by awarding us a patent for our groundbreaking algorithm. This achievement is not just a milestone for our company but a testament to the relentless pursuit of excellence and innovation that drives us every day.

“Our patented technology stands at the forefront of the AI Video industry, setting new standards and pushing the boundaries of what’s possible. This patent underscores our commitment to delivering cutting-edge solutions that address real-world challenges. We are excited about the opportunities this patent opens up for us and our clients, and we remain committed to leveraging our technology to make a significant impact.”