Case Study: Finland’s MTV Oy Leverages NVIDIA’s AI Solutions to Elevate Viewer Experiences, Content Monetization in Broadcast

MTV Oy (MTV) is Finland’s leading commercial television broadcasting and video content company. MTV’s media group includes free TV channels (MTV3, Sub, AVA), an advertising-based video on-demand service called mtv service (mtv-palvelu), and MTV News (MTV Uutiset) that shares the latest stories and news on multiple channels. In addition to the free streaming service, MTV is responsible for the content of a subscription video on demand service called C More.

Valossa offers AI software solutions for video analysis and content intelligence. Valossa’s key products can accelerate video promotion with AI-generated content and metadata, profile videos for contextual advertising, ad targeting and brand safety, and analyze video for sensitive content. The company uses images, video content, speech, person and text content recognition — helping clients deliver content in much more convenient, efficient ways. Using AI, powered by NVIDIA GPUs and SDKs, MTV and Valossa are enhancing audience experiences and content consumption.

Challenge
MTV is part of Telia Company, one of the biggest Nordic TV companies, which is leading the way for the development of the media industry in Finland. To be the market leader in the industry, MTV looks for innovative ways to bring content to life while enhancing the user experience for both viewers and advertisers. MTV aims to:

• Improve content discoverability and recommendations
• Enhance viewer engagement and viewer retention
• And better manage the amount of content and media libraries they own.

In looking for modern, cost-efficient solutions to improve current workflows, MTV chose to explore AI solutions. When MTV started AI development, the team had meetings and test projects with several AI tech providers. Valossa stood out with its great knowledge of AI technology, ways of working, and flexibility. MTV saw an opportunity to succeed in competition against other big OTT and news service providers, both locally and globally, with Valossa’s AI solutions.

Solution
With Valossa AI software, MTV can automate the extraction of essential details that define the main content of videos and other online media. This helps the team enhance operations by improving content visibility in search engines, and assisting the content creation teams through automated categorization and tagging. The AI also analyzes video content and uses the data to automatically generate engaging, promotional video clips.

Valossa AI software is designed to work with a variety of NVIDIA GPUs to meet the diverse needs of customers. The video analysis engine benefits from a broad set of GPU architectures — the more recent the architecture, the better computational benefits are obtained for inferencing and training the AI.

The Valossa AI video analysis cluster is built uniquely to enable hybrid cloud processing in public clouds and data centers. The AI software works with most NVIDIA GPU-powered X86 nodes across public cloud infrastructures. Valossa is also using NVIDIA GPU data center computing and workstations for development and training models with multi-GPU configurations.

For this case with MTV, Valossa is using the TensorRT SDK and CUDA Toolkit to give the AI software more flexibility to switch between GPU architectures. To successfully analyze data such as images, video and speech, MTV requires complex AI software that can deliver detailed results. With NVIDIA products, Valossa built a video analysis stack that can extract various elements, including the smallest details, to create rich metadata that includes contextual video topics and categories, speech keywords, audiovisual tags, celebrity appearances and more.

To read more of the case study, click HERE.