Challenges for Video Analytics

Video analytics seems awesome… so why is it not growing faster?

Video analytics is a trending topic. We have seen that its applications and benefits are endless, and no doubt in the near future we will see many derivations of this technology. But nowadays, what is holding back the growth of video analytics?

This technology is facing mainly the following problems: 

• Accuracy - The system is still inaccurate. This means it still has a lot of errors and raises false alerts and needs maintenance. 
• Human Intervention - It needs human intervention or monitoring, which increases the cost of the system overall. 
• Privacy and Security - Another crucial issue is the growing concern about privacy and the protection of civil rights. The installation of video analytics tools massively will necessarily imply that everyone is seeing anywhere at any time. This is scary. Who is in charge of this information? Who will use it and for what purpose? This could turn into a huge Orwellian Big Brother. Some Hollywood movies are already making us aware of this… Legislation always tends to trail a bit behind industry. The legislative reality also differs from country to country in terms of the degree of safety and guarantees they offer when treating data and transferring data, which makes things even more complicated and confusing.

Apart from just these three main, there are also Environmental and Computational Factors that can affect the performance of Video Analytics. This page can give you a lot more detail on what these are and how it can affect the performance.

Wow...Is it possible to get around this?

Of course it is possible. Eventually, as these issues are taken care of, the costs will be reduced and the technology will become more affordable. As of now, the principal users of the tools  are public administrations, who can pay the price of using this services and also overcome legislative issues, while private ones, more concerned with cost and profit, need to think twice before making use of it.

Have a good weekend and don't forget to come back and visit us next week!

Messi vs Ronaldo.....using VCA to determine who is actually better.

Messi or Ronaldo?  Who’s better? Can it be answered? Yes!

Many have asked the question above, and there has never been a common consensus.  Goals scored, number of assists or individual/ team trophies won do not always dictate the effectiveness of a player has on the game of soccer (football).  However, what if we said that there is a way to quantitatively answer this question, and without using goals scored?  Would you believe it?  Well you should, because the technology does exist.

So how does it work? 

There is company called Prozone Sports (www.prozonesports.com) that does exactly what we speak of. Prozone sports offers services that include post-match analysis, opposition analysis, player recruitment, live performance, referee, and research & consultancy.  It’s unbiased and nonobjective.   It makes transforms all of the elements of the game, including the ball and the players, into “data points” and then uses complex algorithms to analyze the data. 

Prozone sports uses 8-12 high-tech cameras positioned around the stadium to track players movements. Combined with complex software Prozone provides performance analysis based on physical, tactical, and technical components of each player.   Prozone’s products, which are several, are customizable and flexible so that it can fit to each coaches/teams preferences.

Here is a quick video of the technology Prozone Provides - Prozone Sports Ltd - PROZONE3.mpg

So who uses it?

Many of the top clubs around the world use this type of VCA.   These include Manchester United, Real Madrid, Arsenal, and multiple national teams (England and USA).

So who’s ACTUALLY better, Messi or Ronaldo?

We are not sure.  You are going to have to ask employees of Prozone if they have tried to compare the two.   We are just avid bystanders who are fascinated by the idea of combining technology and sports together, and then seeing the results.  

Its no longer Observation....its Observision!

What do you mean?

Have you ever had the opportunity to observe what a Closed-Circuit Television (CCTV) monitoring station looks like? Take it from us, its glum. Imagine being in a dark room with up to 50 screens and you have to monitor all of them. That sounds bad? Well, it gets worse. Remember, that in most of the 50 screens for most of the time....nothing ever happens. And also, you have no clue what you are looking for and what is about to happen. No wonder no child has ever said that he wanted to grow up and be a CCTV monitoring man....

Take a look at the below images to get an idea what life is like in a CCTV monitoring room.

But, that's all about to change though....

Let me guess.....Video Analytics again?

You guessed right! Yes, it is indeed Video Analytics. Using a network of cameras, we can monitor all that's going on. If something fishy is happening, the computers can even zoom in and take a good look and keep the video for later recovery. And in case something really really fishy is going on, it can even send you a message telling you about it so that you can come in and personally intervene.

And the best thing about it all? No human being needs to be sat anywhere looking at 50 screens waiting for something to happen when he could be spending time with his family, watching his local football team, or enjoying a quiet pint at his neighbourhood pub.

No way! What about security? Can you trust a machine?

The question is, can you trust a human? Human capital is limited. Human potential is also limited. Think about it, has CCTV monitoring by humans ever prevented a serious crime? It may have helped in later investigations, but it has never helped prevent one as far as we know. For example, of the 100 people entering the London metro every minute, its impossible for one (or even five) people to continually scan them for any dangerous faces, etc. But using Video Analytics, we can train computers to do the job for us, and they can do it a lot better than we can.

OK, I get it. So how does it all work?

To avoid boring you, this time we have a video form a company called Observision to do the talking and explaining for us. Observision is a Malaysian ICT solutions company, and are one of the pioneers of the application of Video Analytics to the security industry. Watch this and you'll be totally convinced by the awesome potential. I'm sure a lot of CCTV monitoring people will be very excited too when they get their hands on this technology...

We  hope you enjoyed the video and enjoyed learning more about Video Analytics. We'll be back with more fun and innovative stuff for you to enjoy next week!

Video Analytics and Facial Recognition

After this week of holidays and having presented Eye Tracking, we’re going to speak about Facial Recognition. Facial Recognition is a computer system that allows automatic identification of a person from a picture or a video frame (for example). Nowadays, it is easy to detect a face in a picture or video, the tools are performing from the moment it can detect a face on the camera or video. The smartphones, webcams and digital cameras now incorporate facial recognition as a tool for autofocus.

So what could be some of the applications of Face Recognition?

1. Police research to identify a theft recorded by a surveillance camera (close circuit television also focuses on the comportment recognition thanks to the steps analysis or suspect paths in a public area)
2.  Access control secured by a robot
3. And so on

In fact, you certainly already know about the topic via Facebook. The company had to abandon facial recognition in Europe; detecting people’s faces on pictures could put their private lives in jeopardy. But in the US, this technology keeps going: facial recognition is used by government agencies (FBI among others) and private companies. Furthermore, this technology should soon be used by the military industry. The US Air Force could have up to 30,000 drones in 2030 equipped with facial or biometric recognition allowing them to identify individuals depending on their size, age, kind and skin color. As a concrete example, the US used facial identification technologies to identify Osama Bin Laden.

So can you explain how it works? 

 

If you’re familiar with the TV show “Las Vegas” you undoubtedly have already seen it in application. Simply explained, you first need an image or video that is then sent to a software to scan it. The software will detect the position of the eyes to proceed to an alignment. In function of this alignment, a record of various characteristic points of the face is performed (nose, shape of the chin and so on). A geometric design is then saved and will serve as a base on which the researches will be performed.

So, are there any commercial usages of this? All the examples so far are military... 

Of course. The potential marketing and financial implications are pretty obvious. An Italian firm that designs mannequins has developed a model called “EyeSee” which is equipped with facial recognition cameras that record consumer behaviour and classifies them by age and gender. This will provide a better understanding of the client and how to efficiently target him. This reminds us of a passage from “Minority Report”.

Wow! This is quite fantastic....

Indeed it is! We shall leave you with one last piece of trivia. You may think this is pure science fiction but are you planning on going to Brazil for the next World Cup? If yes, be ready to be recorded by the “Robocop Glasses” that are equipped with a camera capable of recording 400 images per second and compare them with a numerical database of 13 million photos.
So if you were planning on doing something naughty in Rio, think twice! 

We hope you enjoyed this week's edition of this blog. Next week, we promise to bring you more action packed and exciting applications of the Video Analytics technology!