Sensara.tv launches EVO: Do It Yourself video content monitoring

In the last two years, at sensara.tv, we've had the opportunity to develop and deploy some amazing video intelligence products to tier 1 customers across the world. Many came with different problems - but they also appeared to need the same foundations. The need to understand and document what is happening in video is universal. Customers are looking to do this at many different locations - in a headend, in a consumer home, in the cloud etc. Many have asked for an affordable and an easy to use product - not long enterprise grade, expensive projects. We've put our learnings together, and extracted a beautiful product. We are talking about this publicly only today - although this product has been out there in spirit in the field solving several mission critical problems already.

We present EVO, an Edge Video Observer. An AI-powered, NoCode, Self serve, Secure, Extremely affordable, Cloud and knowledge graph supported - video content monitoring device. It's simple. Pass your video feed to this device, ask it to look for certain things - and lo, a rich analytics stream - along with visual proof of observations gets recorded and uploaded to the cloud.

We are currently taking pre-orders for the device. Write to us at [email protected] , with your use case. We can take it from there. Thousands of these devices are already on the way to the field, serving critical applications.

Video content monitoring has several applications

Ad Spot Tracking:

Video Advertisers want to study their ad placements, and verify if the publishers have indeed carried them. A hard problem in the broadcast world only got harder in the digital world - with absolutely no visual proof being available to advertising spenders.

On the other side, several networks and broadcasters - especially the small and upcoming ones need to present proof of carriage to their customers.

Ad spot tracking has been an expensive affair so far - involving complex setups and machinery, either on premise or even in the cloud. Not any more.

Music Monitoring:

Creating music is hard enough. Protecting it from piracy and misuse is even harder. Billions of dollars are lost every year since plagiarism goes unchecked in the video industry. Record labels want to ensure their tracks are used legitimately and want an early alarm system so they can track down and warn offenders. This process requires them to submit proof of violation.

The EVO makes this easy. Record labels can pick and choose any playlist of content from any source - broadcast or even OTT - and feed them continuously to the EVO for monitoring. The EVO can index thousands of titles easily and look for violations.

People monitoring:

Video influences the minds of people. Elections, purchase behaviour, investing behaviour, fashion, food, etc. are all influenced by people that come on Television. But who occupied the screen for how long? EVO makes this simple to find out. Just feed any channel of your choice and ask it to monitor for people's screen presence.

The EVO is supported by a Knowledge Graph of visual features. Sensara.tv has been silently building this for years now - a Knowledge Graph of who comes on television and how to identify them. Identifying a lot of people got easier over time. EVO just downloads these features onto the device and looks for them in the video. The unknown people are clustered and asked to be labelled.

So, how does the EVO work?

The EVO, an acronym for Edge Video Observer, is a video input processor - on a single-board computer, that is specially built to recognise audio-visual features, and report occurrences along with visual proof (a screenshot or a clip) to the cloud. Every observed event is sent securely to the cloud and a timeline of video events is created. When you want the EVO to look for something, you just come to the EVO cloud and tell it what to look for - through examples.

The nice thing about the EVO, is that it is both Cloud SaaS and the Edge. Moving videos to the cloud is expensive. Analysing them in the cloud is even more expensive. The edge is where videos are easily available. But the cloud is great for control and for consolidating results. Instead of moving videos to the cloud, we move the smartness near the video. Numbers and events are moved to the cloud.

This is sophisticated distributed systems work. We send the recognition models to the edge, and keep them current - when the specifications change. Different EVOs get different configurations.

The EVO is supported by one of the largest knowledge graphs of visual features. Over the years, Sensara.tv has learnt how people look like, their names, occupations etc - based on their occurrence on Television. Sensara.tv has also learnt about thousands of brands that go around in TV. These visual features have all been pre-trained. So, when you are trying to observe something in a video - chances are - we've got you covered already.

EVO Cloud loves APIs. We have APIs that allow you to programmatically and securely upload - like ad creatives, people faces, songs etc. We also have APIs for you to take out recognized video events, and visual proofs (as png or mp4). You can ask different EVOs to do different things.

EVO also integrates well with some of Sensara.tv's enterprise offerings - like ad mining, people mining etc. This means you can simply come to EVO cloud and ask for an entire category of ads from different brands to be monitored. If we already know something, why should we get you to tell it again?

The EVO is already well tested

Technology at work inside the EVO has been spotting ads for 5 years now, building India's largest ad database ever. The same technology has been at work spotting ads in real time and swapping them out in Nepal and Bangladesh. More than 20000 songs are being checked for everyday across 30+ broadcast channels for copyright violation. Proof of flight records for ads are being generated across house channels. We are opening this up - so we can figure many more use cases and serve many more customers.

The EVO should open your exploring mind

Here's a simple way of using the EVO to setup a TV Attention Meter. Suppose you want to know how a home reacts when they watch something on TV.

Many of these use cases have been multi year custom hardware projects - that now become No Code. Customers can experiment on a small number of deployments and then scale up. Do you see why we are so excited?

EVO Configurations

So far, for the use cases we have been considering, we perceive video inputs to be made available through HDMI, IP (UDP/Mpeg-ts) or SDI. We've created different configurations for these. Different types of use cases need different memory profiles as well.

The Pricing is the best part

The best part of the edge is that you've already paid up for the compute. So, what you pay for is for the intelligence and the service provided. Costs are brought down drastically - compared to other cloud offerings for the same amount of work.

Do we see you reaching out to us? :) Let's talk about your use case and how many devices you'd need. We are taking pre-orders. Write to us at [email protected]