Top 3 reasons why computer vision is changing the asset management game
The expression ‘content is king’ is well known in media and entertainment. But the truth for many companies is that ‘content is complicated’. How do you manage all your digital assets including thousands of hours of video, and how do you maximize the efficiency of workflows that support the complex business of sourcing and editing content?
Many organizations have deployed digital asset management (DAM) systems that go some way to solving these challenges. But as the volume of video increases and new market entrants put a squeeze on revenues, the pressure to improve content retrieval and workflow efficiency is greater than ever.
All these challenges are close to the heart of Iain Churchill-Coleman, Product Manager at EditShare, a supplier of network shared storage and media management solutions to media and entertainment customers worldwide. These organizations use EditShare’s software to bring ideas to life: from advertising and corporate videos to traditional broadcasting and new media.
“EditShare exists for the simple reason to help organizations tell exciting stories and to provide them with the tools to publish high quality video to their audiences. As the sector is transformed by new media and platforms, our mission is to help our customers evolve to meet the challenges in this fast-moving landscape,” he says.
One of EditShare’s most widely used products is FLOW which enables EditShare’s customers to manage millions of digital assets no matter how and where they are stored, including cloud and on-premise solutions. Primarily aimed at online editors and producers, it includes sophisticated workflow features that enable teams worldwide to communicate and collaborate easily.
“One of FLOW’s greatest benefits is its flexible storage and adaptable workflow model,” says Churchill-Coleman. “You can scale as and when you need - from two users up to hundreds of team members.”
Until recently, when digital files arrived in the FLOW environment, clients would spend hours or even days manually tagging hundreds of hours of video before any footage reached the craft edit suite. This was one aspect of the solution that Churchill-Coleman and the EditShare team were keen to address.
This time-consuming process included trying to find specific people in a shot, locate a venue or identify a product for an advertising clip. And because it was a manual process, there was always a risk of human error with some content going untagged and unindexed.
During this time, EditShare was exploring the possibilities of artificial intelligence and computer vision to solve this problem by automating the tagging process. “We started researching the marketplace and going through our contacts in the pursuit of a viable AI solution that we could integrate with our technology,” says Churchill-Coleman.
Having researched the market and then meeting with the Mobius Labs team, EditShare selected the Berlin-based start-up’s AI software to power a new feature called FLOW AI. This takes advantage of Mobius’s Superhuman Vision™ technology to tag objects, people, locations and even emotions. It also generates time stamped audio transcriptions so that editors can search for a clip by entering a line of dialogue or commentary.
“Automation is critical to the EditShare offering,” says Churchill-Coleman. “Thanks to Mobius and FLOW AI, content can be tagged when it is first ingested, or you can run the software to go through your archives and tag historical files just as quickly and efficiently.”
The FLOW AI keywording platform can instantly recognize more than 5,000 objects, emotions, actions, and any people added to the database. Rapid indexing dramatically reduces the hours spent on manually indexing content, thereby saving valuable time, effort and expense. Now, for example, content can be captured after the day’s shoot, delivered to a facility, and indexed and delivered to editing suites, without the need for a technical operator.
Retrieving archived content is equally straightforward. Production teams can find the shot they need by simply typing keywords into the FLOW AI search engine and the clips are returned instantly.
The response from EditShare customers and partners has been universally positive. This includes Imagine Communications, whose solutions, in conjunction with EditShare’s FLOW, enable news organizations to publish stories from reporters in the field and from their archives. “Take the example of an obituary,” says Churchill-Coleman. “Once you have tagged your archive, you can find and then stitch together clips of a celebrity or a historical figure much faster than ever before.”
Mobius, as part of FLOW AI, also enables EditShare to widen its appeal to international customers. “Localizing content can be a big challenge. But with FLOW AI you can flag problematic segments such as nudity or smoking, and filter them out of the search results.”
EditShare has also been impressed by Mobius Labs’ software roadmap and its willingness to accommodate feedback from customers. “The team at Mobius are amazing from developers to the sales team. They don’t just wait for our feedback; they are always coming to us with new ideas.”
As a result, Churchill-Coleman is looking forward to further collaboration with Mobius as the AI landscape expands, driven by faster CPUs and greater volumes of data. “Mobius has amazing computer algorithms that not only save time but offer more accurate and richer results. Here at EditShare, we are looking forward to further collaboration as Mobius Labs’ computer vision software evolves,” he says.