“Human moderators simply can’t keep up with the current content overflow,” says Valossa CEO.
HELSINKI, FINLAND – We are spoiled with video content. According to Cisco Visual Networking Index, nearly 80% of the IP traffic is already video and the portion is growing 26% annually. It means that if an individual would watch all the uploaded content of a single month in 2021, it would take five million years.
The ever-increasing volume of content is particularly challenging for online video platforms, which are distributing content in high velocity and pressured to moderate inappropriate or unlawful content to protect the community and children. This means that companies like Facebook and YouTube have thousands of employees for content monitoring.
Broadcasters and traditional media companies are also struggling with the problem of content management and all the manual labor required to ensure content compliance with for example age restrictions and broadcast times, according to national laws and regulations.
Most of the moderation is done manually by visual review. However, sophisticated machine learning is becoming available. Several providers are using image recognition techniques to moderate pictures using modern cognitive computing. But video content compliance and moderation require solutions that are more complex than recognizing objects from a single picture.
Valossa AI is a video recognition solution that uses a more comprehensive video recognition approach to extract actionable content intelligence. It watches videos at superhuman speed extracting sequences of speech, sounds, and visual objects, and seeking events that are relevant for content reviewing. Obtained results are summarized into content metadata and available as visual, interactive reports for content inspection. The report contains categorized, time-based segments of potentially inappropriate content. Alongside of the compliance report, AI lists most important people from face recognition and profiles content based on themes coming out of the audio-visual context analysis.
What makes Valossa AI solution stand out from the competition and well-known rivals is the audio-visual coverage and accuracy proven by benchmark tests. Compliance-wise, the AI can recognize a broad variety of potentially inappropriate concepts in nuanced levels of nudity, sexuality, and violence. It also spots events and objects depicting substance use and disasters, utterances of bad language, and gives summaries on how strongly they are represented in the content.
The operator may upload the material for processing, let the AI do the heavy lifting, view the reports and get sub-second accurate timestamps when the flagged content starts and stops.
Valossa’s solution benefits the media industry processes broadly because the API can be integrated into a variety of customer workflow tools and platforms. In addition, Valossa develops AI-based content inspection tool suite, Video Insight Tools, to let video businesses inspect, profile and discover content with the power of modern AI techniques, and even train it further to better fit their content management purposes. Video Insight Tools are available for a free trial at valossa.com.
“Human moderators simply can’t keep up with the current content overflow. Whether it’s offline video or live-stream, Valossa AI analyzes and provides insight at superhuman speed.” -Mika, CEO.
The roots of the video AI company Valossa are in one of Europe’s leading computer science and AI labs at the University of Oulu in Finland. During the years in the university research labs, the team had seen how the world moved beyond text, with video playing a greater role in our daily communication. This spawned an idea to use technology to automatically understand what we are viewing and apply the latest techniques to meet the emerging needs of content producers globally. Since their founding in 2015 Valossa has built next-generation video intelligence tools for the benefit of the businesses working with video.
Valossa provides unparalleled video recognition and content intelligence capabilities to capture new value of video. Valossa AI is a world-leading technology platform that enables a completely new generation of analytics and content profiling tools bringing forward new ways of monetizing video.
Article originally featured in Analytics Insight / November 2018 issue.
Online version can be found at https://www.analyticsinsight.net/managing-content-overflow-video-recognition-technology/
Marketing & Communications Manager