7 important items to consider before investing in AI video analytics softwares

Lost in the jungle of neural networks, metadata, pose-estimation technology, deployment models, computer vision algorithms, software architecture, and data sheets?

We are here to help.

There are a lot of products on the market these days, and let’s be honest – it can get quite complicated sometimes. So it’s understandable that the buying process can be overwhelming for someone who is looking to invest in video analytics softwares for the first time.

To put you on the right track, we’ve created a list of must-dos and please-don’ts to think about before getting started, focusing on the following subjects:

1. Cyber security
2. Integrations
3. Functionality
4. Scalability
5. Deployment model
6. Compliance
7. Support

1. Double-check that the system is secure against cyber threats

This is extremely important for two main reasons:

1. You cannot, as a security company, have a data breach. It would have serious, negative repercussions on your brand.
2. Even a small breach can result in huge fines that can cripple your company.

The key is investing in video analytics softwares that boosts the performance of your security infrastructure without exposing your business to cyber threats.

Make sure that safety mechanisms, like secure authentication and access features, are in place to prevent unauthorized users from accessing your system.

Typical entry points for hackers are back doors embedded in the software that are used for customer support and troubleshooting. Be sure to ask about these and the steps that have been taken to stop hackers from using them to bypass security.

Shortcut to success: Take advantage of the best-in-class security practices from the major cloud providers.

2. Make sure that it’s easy to integrate with your existing hardware and software

If you get a headache from all the extra things you need to do, or buy, to get the system up and running – something’s wrong.

First and foremost, look for a system that works out of the box. It should work immediately without the need for further optimization or training.

Then you want to know if it’s camera agnostic (compatible with any brand or type of camera)?

Can it integrate with your video management system (VMS), or do you need to buy new hardware for it to work?

How much storage space and internet speed do you need to run the system – and we don’t mean just barely, but at optimal settings?

You absolutely do not want to find out that you need to spend a ton of extra cash on new IT equipment because the bandwidth requirements of your new system are way too high for you to even use it – we promise you that it isn’t a nice surprise.

You should particularly note the quantity of data that is sent between your cameras and analytics device (“edge device”). This is typically where you will find the biggest bandwidth burden.

Ask how vendors have worked to reduce this burden by being selective about the type and quantity of metadata that is sent between these two devices. Because if you’re always sending full, data-heavy, video streams – think 720p-1080p at 30fps – you might run into problems.

3. Which functions do you need, and which would be nice to have?

There are a lot of options out there. So, before you even get started, make a list of your must-have’s and nice-to-haves before you go shopping.

Do you need real-time alarms or forensic search? VMS included or stand-alone? Maybe you need data insights, such as counter-flow traffic or traffic statistics?

Vendors should address these needs and highlight important aspects that you haven’t even considered.

Important note! Once you’ve found a product that offers all the functionality you need – double, triple, and quadruple check that it keeps false alarm rates to a minimum!

Don’t get caught in the trap of a fancy-looking technology that will plague you with endless false alarms. Ask to test the product before you buy it and check to make sure it has a proven track record of low false-alarm rates.

4. Invest in a product that can grow together with your business.

How easily can you add new cameras or servers to the system? This should be simple.

Do you want to add new analytics functions, like slip and fall, to your network? This shouldn’t be a problem.

Whether you have a security infrastructure with 10 cameras at one site, or 10 000 cameras spread over multiple sites, implementing a good AI system should be seamless.

Your business may change over time, but your analytics system should have no problem adapting to suit any shape or size it takes.

5. Look for a deployment model that suits you

From installation, to functionality, to pricing models – you want a system that adapts to suit your particular needs. Depending on the size of your company, both small or large-scale deployments should be possible.

Do you prefer an on-premise or cloud-based installation? Maybe a hybrid model is the best fit? The answers to these questions largely depend on how much responsibility you want to take for the maintenance of your IT.

Look for vendors that offer different service packages. If you don’t need certain functions, you shouldn’t need to pay for them.

6. Don’t forget compliance with data-privacy laws!

Please, do not overlook this.

If you do, you could end up with huge fines for jeopardizing the personal data of subjects recorded on camera – and we do mean huge.

Guarantee that mechanisms are in place to minimise your exposure to potential data breaches, such as anonymization features and flexible permissions for different user groups.

These features facilitate your compliance with local data-protection laws, like GDPR. They allow certain members of your organisation, or the authorities, to access video footage in the event of, say, a police investigation. However, the video data would remain anonymized, or inaccessible, to other parties.

7. Get the support you need to make the most out of the product

Look for a vendor that will continue to support you long after the sale.

Will they provide you with 24/7 network and camera-health monitoring? Will they alert you if any of your cameras have been tampered with?

Is there a support email available with timely responses? A phone number?

One last thing! If you decide on a hybrid or on-premise deployment, ask how your system can be kept up to date with the latest software and features.

You don’t want to waste valuable time and resources on sub-par analytics. But by following the tips above, we’re confident that you will land on a quality system.

Just remember, what it boils down to is this:

The right AI video analytics softwares should make your life easier, your cameras more effective, and provide your business with a tangible ROI.

Read more about Irisity’s solutions for AI Video Analytics >>