From POS Text Overlay to Integrated Video Analytics

The evolution of transaction integration
By Chris Beaudoin 3 years agoNo Comments

Throughout the 70s and 80s, finding any evidence in surveillance video was a very time-consuming process of rewinding the tapes. Since then, the security industry has seen some incredible advancements in technology and POS Text Overlay is one of those. But of course it doesn’t end there: Text Overlay is just the beginning of Transaction Integration.

A Brief History

In the 1980s, TVs with on-screen display (OSD) started to be more common. OSD allows an image to be superimposed on a screen to display information such as the volume or channel number. Soon after, it became apparent that there was a huge opportunity to use this technology in the security industry.

The idea was to superimpose transactional data from point-of-sale (POS) or ATM systems on top of the video feed. To do this, Video Text Inserter devices emerged, also called POS Text Overlay or Video Serial Interfacing. The devices have the ability to trigger VCR recording so that video is only captured when a transaction is taking place. Later, DVRs (digital video recorders) were developed with built-in text overlay capabilities.

The Value of Text Overlay

Auditing transactions used to mean looking at two completely separate systems (the POS or ATM machine and the security cameras) and trying to manually match a receipt to a few minutes of footage. This is known as time and location search and it limits you to just reactive investigations since you have to have a suspicion that fraud took place at a certain time and place before you look for the evidence. Otherwise there is just too much data to review.

Auditing transactions before Text Overlay

Auditing transactions with Text Overlay

  • Find out about suspicious transactions
  • Get the time of targeted transactions
  • Pull up transactions in the POS records
  • On video system, search the video by the time
  • Watch the recorded video
  • Compare video to transactions
  • Find out about suspicious transactions
  • Get the time of targeted transactions
  • Search the video by the time
  • Review the video with transaction information

When text overlay was introduced in the security industry, it was a game changer for many business owners and fraud investigators. It allowed the person reviewing the footage to have a visual comparison on the same screen of what items are being rung into the register and what items were actually handled by the employee and customer.

This made the auditing process faster, smoother and more effective at catching fraud. Text overlay is still widely used and an integral part of loss prevention.

text overlay

 

Text Overlay is great, but it has its limits

While the technology helped people to detect internal and external fraud in a completely new way, it has its limits. If you don’t have a DVR with built-in text overlay, you will have to buy the additional hardware device and connect it to the DVR. While this in itself isn’t a major problem, it can become very expensive. Some of these devices cost upwards of $500 which can really add up if you are monitoring multiple locations.

The second issue relates to the ability to search and filter the text overlay information. Transactions are by definition a structured data source, just like a spreadsheet. It’s easy to search for a specific value. Video and images however are unstructured; the only parameter that video has is the time of recording.

With text overlay, you are taking a structured data source (transactions) and essentially turning it into an image which is layered on top of video frames (both unstructured and unsearchable). As a result, the auditing process still relies on time and location search.

How Exception Reporting made it better

Luckily, most DVRs and text overlay devices today have addressed the second issue by keeping the transactional data as text rather than converting it into an image. With this change, you can search the data by product name, price, or any other text query.

But the real improvement came with Exception Reporting. Once the text data was searchable, creating alerts and reports became a possibility. With Exception Reporting, a transaction that meets preset filters or conditions is flagged and then triggers an action such as sending an alert or being added to a report.

Some examples of filters and conditions:

  • transactions marked with “void”
  • transactions less/greater than or equal to a certain amount
  • transactions with a discount

Exception Reporting replaced the majority of time and location searches that were necessary. Instead of conducting periodical audits or waiting to learn about potential fraud before investigating, you can be notified about high risk transactions as they happen. This was a major step towards proactive investigating.

Auditing transactions with Exception Reporting

  • System flags suspicious transactions
  • Each transaction along with video footage is organized in a report/creates an alert
  • Review the video with transaction information

Exception Reporting made the POS system into an intelligent source of data. But the truth is that an image is worth a thousand words, while a receipt is only worth so many. So while text overlay is incredibly valuable, it is literally laying on top of an even greater data source that hasn’t been exploited. This is where video analytics comes in.

Today: Transaction Integration and Video Analytics

Video is the most reliable and context-rich source of information. In fact, we trust it more than we trust our own eyes and memories. This isn’t surprising as our memory often plays tricks on us. Just look at the sports industry where replays are essential to help commentators and referees make accurate judgements about a play.

If text overlay and Exception Reporting make your POS system intelligent, imagine what would happen if your surveillance video was also intelligent. What if both systems were working together? This is the goal of video analytics and transaction integration.

But correlating different data sources is not as simple as overlapping transactions on top of video. The two data sources are not actually speaking to each other, they are just displayed together. At Solink we’ve developed Integrated Video Analytics technologies so that our system can pull information from transactions, video and any other data sources and determine if each source is telling the same story or if something isn’t right.

For instance, with video analytics, our system can determine that a customer was not present during a transaction and flag this event as an exception. Knowing that an employee opened the cash register when no customer was present helps you to be proactive in detecting internal fraud. There are certain situations like this one which cannot be detected by a POS Exception Reporting system and this is why it’s important to use video analytics.

Moving beyond loss prevention, video analytics also lets you make connections between the transactions that took place and how many people entered the store or branch, how long they shopped for, if they stopped to look at a promotional display and more.

Auditing transactions with Solink Transaction Integration

  • Every transaction automatically creates a details page (shown below)
  • These can be searched, filtered and flagged in the same way as POS Exception Reporting
  • Video analytics allows additional exceptions to be detected

text overlay

Takeaway

Text overlay and exception reporting paved the path towards smart POS systems. Now we are using integrated video analytics to create surveillance systems with intelligent capabilities. We want to help people realize that video holds a tremendous amount of information about their business. By merging the physical security and data analytics industries, we can create better ways to audit, detect fraud, improve operations and deliver value to customers.

To learn more about Solink Transaction Integration, get in touch with us.

Categories:
  Finance, Retail
})(jQuery);