Storyblocks, an unlimited, subscription-based stock media platform with video creation and workflow functionality, has officially announced the launch of several new search enhancements that leverage AI to conceive an elevated user experience and streamlined creative workflows for Storyblocks’ customers. Talk about the development on a slightly deeper level, we begin from the promise to deliver faster, more relevant, and more representative results in search queries. Here, the core idea is to prioritize fresher content in search results through a proprietary algorithm that evaluates the performance of new content every four hours. Complimenting the same is a particular discovery, which revealed that recently uploaded content can boost download rates by up to 60%, in comparison to searches without this prioritization. Next up, we have the prospect of employing an in-house machine learning model based on the BERT framework to help Storyblocks’ platform rapidly analyze how users engage with content, and therefore, refine search result rankings in near real-time. As for the company’s bid to deliver more representative results, it carries Storyblocks’ practical objective to feature models from underrepresented backgrounds across, at least, a quarter of its total library content.
Moving on, Storyblocks has further made a decision to deploy AI when looking for thumbnails that will show up in video search results. The stated move followed-up on another recent discovery where it was figured that videos showcasing AI-curated thumbnails experienced heightened engagement levels. Anyway, we touched on the platform’s resolve to always show relevant results, it builds upon that by banking upon OpenAI’s large language model to carve out a setup that returns accurate results even for flawed search parameters. On top of it, using metadata enhancements and Storyblocks’ computer vision technology, the platform is now looking to aggregate pertinent search results, recommend alternative scenes, and surface additional clips featuring specific individuals.
“At Storyblocks, we’re harnessing the power of AI to revolutionize the way creators discover and use content,” said Lucy Huang, senior vice president of product and engineering at Storyblocks. “By delivering faster, more relevant and more representative search results, we’re not just streamlining creative workflows – we’re empowering our global community of artists and customers to create inclusive and impactful work. Our commitment to curating a diverse and top-quality library, combined with cutting-edge technology, sets us apart in the stock media industry.”
Rounding up highlights is introduction of cross-asset recommendations. The creative assets in play here include videos, images and audio tracks. To offer some details on the way this technology works, Storyblocks would apply its computer vision technology to analyze, let’s say, music tracks by rendering songs as spectrograms. Once converted, these spectrograms allow the platform to identify and recommend tracks that are significantly similar.
Founded in 2011, Storyblocks has risen up on the back of its ability to produce better and faster videos. This it does through a library containing 2.6 million assets, stretched across high-quality video, audio, and imagery. Aiding the case of these assets even further is the fact that they are crafted by more than 700 highly accomplished artists and creators from all over the globe.
“Our team has relied on Storyblocks for several years now to create video content at a rapid pace,” said Nolan Nichols, manager of video production at the U.S. Travel Association. “We’re excited that Storyblocks is leveraging AI more to help customers like us find the content we need even quicker than before.”