Cognitive Services
Evolve your business model with machine learning
What are cognitive services? - Definition
Recognizing content and objects and understanding a spoken language have set us apart from computer capabilities since the dawn of computers. Although we cannot crunch numbers at the same speed as a computer, we are aware of the story in the film, we can appreciate and describe the contents of the media files, we can understand languages, and we recognize actors. These skills are examples of cognitive abilities - from the word "recognize".
Wouldn't it be great if we could combine these cognitive abilities with the nearly unlimited performance in today's computing environments? Well, it's already happening. With the introduction of machine learning software, it is now possible to train algorithms to recognize media contents the same way humans do. The definition of cognitive services refers to software that enables computers to understand or process external information like a human, but with much greater processing power.
Create your cognitive supply chain with cognitive services API
The time has come when we must challenge our perception of what computers can or cannot do. With the development of cognitive machine learning algorithms, CTOs, Content Owners, and technical strategists now have an entirely new set of tools to build the next-generation digital media supply chain.
With the introduction of VidiNet Cognitive Services, VidiCore API and VidiNet became the perfect platform for the job.

Five examples of using cognitive services to achieve positive ROI - Evolve your business model with machine learning today!
Why use the Cognitive Services from Vidispine?
Media companies, news companies, etc., need the capabilities to manage an increasingly larger and more complex stream of video content delivered to them, not only in quantity but also in the growing amount of video formats. Manually sorting through maybe thousands upon thousands of video assets is a waste of both time and resources. By combining cognitive services with the flexible media management of VidiCore, you can easily create and utilize cognitive metadata to your advantage.
Vidispine has extended the VidiCore API with a dedicated cognitive services interface. It is a core services technology framework giving you as a customer direct access to a growing number of partners offering cognitive services on the market, including speech-to-text and image and face recognition cognitive services. In VidiCore API, we have defined a standard structure for the cognitive metadata from different providers. Because of this, customers don’t have to care about how to model and integrate the different metadata results coming back from various providers.
How can cognitive services improve your existing Media Supply Chain?
AI and machine learning exceed human performance in many areas, and more is to come. Research from the University of Oxford predicts that AI will outperform humans in many domains in the next 40 years, such as:
- translating languages by 2024
- driving a truck by 2027
- working in retail by 2031
- writing a bestselling book by 2049
- working as a surgeon by 2053
The researchers believe there is a 50% chance of AI outperforming humans in ALL tasks in 45 years, and that ALL human jobs will be fully automated in 120 years. If this is true or not, we do not dare to speculate. However, we understand where the notion is coming from when looking at the current state of cognitive abilities in AI. Here are a few advantages of using cognitive services API to transform your existing supply chain into a cognitive one:
- Eliminates manual work, speeding up workflows and reducing the risk of human errors.
- Automated creation of detailed and vast amounts of metadata for all your video assets.
- Improved search capabilities, allowing you to perform advanced searches in your media library to find the video assets you need when you need them.
- Automated compliance checking ensures all your videos are following the platform’s guidelines.
- Creating highlights, trailers, and shorter videos automatically, for example, a highlight reel of a specific soccer player’s goals and passes.
Features of cognitive services offered by Vidispine
With the ability to train machines with large numbers of audio-visual data, many new opportunities open up where computers and humans can work side by side. Once you have the information as metadata you are can use this information improve the performance of your media supply chain.
-
Automatic Subtitling
A machine learning algorithm can do the main part of the job, and a human can make the last changes. Depending on the environment, cognitive services for a speech-to-text algorithm can achieve at least 80-90 % accuracy. Imagine the time spent in manual transcribes for subtitling; now cut that time in half… twice. That is how fast automatic subtitling using cognitive services for speech to text can be
-
Face recognition
Being able to train the algorithm to detect faces can prove very useful for tracking actors in a production environment in a ingest process. These types of face cognitive services are, of course, also useful for determining the amount of time a person has participated in or finding all clips in an archive with a specific person.
For example, maybe you want to know what a certain person has said about a certain topic? In such a case you would use a combination of automatic transcription and face recognition.
-
Image recognition
Using cognitive services image recognition to collect data about image contents can be very valuable. For example, it allows you to find all shots from the outside location with a certain actor, or detect a brand's presence in a movie. The current state of the Vidispine Cognitive Services relies on the AWS image recognition algorithm. It tracks all objects available in your video with a confidence value to be used as a factor of accuracy.
Conclusion
The capabilities of AI-based cognitive services are improving, and the number of cognitive services API providers is growing rapidly. Having a MAM system with the technology framework to take care of cognitive metadata from different providers lets the system do the heavy lifting of modeling and integrating metadata results. The VidiNet cognitive services offer an intuitive and efficient way to extract and generate tailored metadata, streamlining the management of your video media assets.