Over the past two decades, the commercialisation of video content has increased exponentially. However, serving and finding quality content in this ever-increasing sea of supply is only becoming more challenging for creators and providers. To a large extent, this is due to the fact that video search and content discovery technologies are still stuck in the last century. Primitive content indexation, based on manually added metadata, makes discovering quality content almost impossible. As a result, creators and providers of quality video fail to reach their intended audiences and are quickly losing viewers to online content providers.
The MoDELS project enables us to commercialise our revolutionary Deep Content Understanding platform that really understands video to a degree that has been unimaginable, up until now. By systematically analysing video and extracting next-generation metadata about every millisecond of content in real time, we will improve the discoverability of content by orders of magnitude, directly influencing your viewer’s experience.
Within the EU project, we have just finalised the below prototypes:
Overlay Detection is now available
Our technology is now able to detect name tags and other text overlays in video. This enables an advanced search on overlay texts such as name tags in video. By enriching the metadata, we are increasing the accuracy and our clients won’t miss a thing.
Improve language models in real time
By automatically updating our language models, our clients are up-to-date with any new word trends. Our technology can now automatically build models on regular basis and provides high accuracy in speech detection.
Logo recognition taken to the next level with our Automatic Cue Classifier
We are now able to detect transparent cues in videos, like channel logos. This helps us providing correct start times for all TV channels we analyse for a highly accurate EPG correction.
A new platform for deployment and scale up
We have standardised the deployment of our platform using state-of-the-art automation and orchestration tools (Terraform, Docker, Kubernetes). The platform is now more scalable, easier and faster to onboard new customers and create new features answering to our customers’ needs.