For English see below

In deze SPAR winkel testen wij de impact van AI op het tegengaan van winkeldiefstal.

Deze SPAR locatie heeft, net als andere winkels, altijd al te maken gehad met diefstal, maar wij zien het de laatste tijd steeds meer toenemen. Om diefstal tegen te gaan, testen wij nu in deze SPAR winkel AI-gedreven videoanalyse.

This SPAR location, like other stores, has always dealt with theft, but we have seen it increase more recently. To combat theft, we are now testing AI-driven video analysis in this SPAR store.

Video surveillance alone falls short

To reduce theft, we have been using various techniques for years, with video surveillance being the most well-known. Unfortunately, this technology is far from foolproof. In practice, it is impossible to monitor video footage throughout the entire day. It is labor-intensive, very costly, impacts customer privacy through live viewing, can cause profiling due to biases of the person watching the footage, and significantly increases the chance of errors. We now want to test whether the use of Video AI Analytics can address the issue of shoplifting. But what exactly is Video AI Analytics?

What is Video AI Analytics?

Video AI Analytics involves the use of smart algorithms to interpret video footage. This algorithm is not a so-called self-learning algorithm but a fixed algorithm that has been tested by the supplier. As a result, the system can automatically recognize what is happening within a certain field of view. The video analysis software distinguishes between different objects and recognizes certain activities or behaviors that trigger a user response or alert. The video analysis software does not focus on people's appearances, does not use facial recognition, but recognizes behaviors and movements commonly associated with shoplifting. Simply put, the software converts video footage into data that can be interpreted. We focus on unusual and abnormal shopping behavior. For example, if someone puts an item from the store into their pocket, authorized store personnel will be alerted.

Real-time alerts

The open nature of a Video AI Analytics system allows for real-time alerts. These alerts are immediately passed on to authorized and trained store personnel, who can respond quickly.

The privacy of our customers is protected through privacy filters. This means that all faces are made unrecognizable and can only be viewed by authorized personnel if an incident has been reported. Footage from the smart cameras is not stored if no unusual or abnormal shopping behavior is detected. If such behavior is detected, the footage is generally kept for 28 days. After the test, we will decide whether we will implement this system in other stores as well. After the test, we will decide whether to implement this system in other stores.

For more information about the use of Video AI Analytics and the associated rights of our customers, we refer you to our Privacy Policy: SPAR | privacyverklaring van SPAR

SPAR Holding, Waalwijk June 2024

Videobewaking alléén schiet tekort
Om diefstal terug te dringen, zetten wij al jarenlang verschillende technieken in, waarvan videobewaking de bekendste is. Helaas is deze technologie lang niet altijd waterdicht. In de praktijk blijkt het namelijk onmogelijk om gedurende de hele dag videobeelden te monitoren. Het is arbeidsintensief, zeer kostbaar, grijpt door het live uitkijken in op de privacy van klanten, kan profiling veroorzaken door vooroordelen van de persoon die de beelden uitkijkt en maakt de kans op fouten aanzienlijk. Nu willen wij testen of de inzet van Video AI Analytics de problematiek rondom winkeldiefstal kan aanpakken. Maar wat is Video AI Analytics precies?

Wat is Video AI Analytics?
Video AI Analytics houdt in dat slimme algoritmen worden ingezet om videobeelden te interpreteren. Dit algoritme is niet een zogenaamd zelflerend algoritme maar het gaat om een vast algoritmen dat door de leverancier getest is. Hierdoor kan het systeem automatisch herkennen wat er binnen een bepaald gezichtsveld plaatsvindt. De videoanalyse-software maakt onderscheid tussen verschillende objecten en herkent bepaalde activiteiten of gedragingen die een gebruikersreactie of een melding triggeren. De video analyse software let dus niet op het uiterlijk van mensen, doet niet aan gezichtsherkenning  maar herkent gedragingen en bewegingen die bij winkeldiefstal veel voorkomen. Simpel gezegd zet de software videobeelden om in data die geïnterpreteerd kan worden. Wij richten ons op afwijkend en abnormaal winkelgedrag. Bijvoorbeeld: als iemand een artikel uit de winkel in zijn of haar zak steekt, wordt het geautoriseerde winkelpersoneel gealarmeerd.

Real-time meldingen
Het open karakter van een Video AI Analytics-systeem maakt het mogelijk om real-time meldingen te ontvangen. Deze meldingen worden direct doorgegeven aan geautoriseerd en opgeleid winkelpersoneel, die snel kunnen ingrijpen.

De privacy van onze klanten wordt beschermd via privacyfilters. Dit betekent dat alle gezichten onherkenbaar worden gemaakt en alleen kunnen worden bekeken door geautoriseerd personeel indien er een melding is gemaakt. De beelden van de slimme camera’s worden niet bewaard als er geen afwijkend of abnormaal winkelgedrag is geregistreerd. Is dat wel zo? Dan worden de beelden in de regel 28 dagen bewaard. Na de test zullen wij beslissen of wij dit systeem ook in andere winkels gaan gebruiken.

Voor meer informatie over het gebruik van Video AI Analytics en de bijbehorende rechten van onze klanten, verwijzen wij u naar onze Privacyverklaring: SPAR | privacyverklaring van SPAR.

SPAR Holding, Waalwijk Juni 202

In this SPAR store, we are testing the impact of AI on preventing shoplifting.

This SPAR location, like other stores, has always dealt with theft, but we have seen it increase more recently. To combat theft, we are now testing AI-driven video analysis in this SPAR store.

Video surveillance alone falls short
To reduce theft, we have been using various techniques for years, with video surveillance being the most well-known. Unfortunately, this technology is far from foolproof. In practice, it is impossible to monitor video footage throughout the entire day. It is labor-intensive, very costly, impacts customer privacy through live viewing, can cause profiling due to biases of the person watching the footage, and significantly increases the chance of errors. We now want to test whether the use of Video AI Analytics can address the issue of shoplifting. But what exactly is Video AI Analytics?

What is Video AI Analytics?
Video AI Analytics involves the use of smart algorithms to interpret video footage. This algorithm is not a so-called self-learning algorithm but a fixed algorithm that has been tested by the supplier. As a result, the system can automatically recognize what is happening within a certain field of view. The video analysis software distinguishes between different objects and recognizes certain activities or behaviors that trigger a user response or alert. The video analysis software does not focus on people's appearances, does not use facial recognition, but recognizes behaviors and movements commonly associated with shoplifting. Simply put, the software converts video footage into data that can be interpreted. We focus on unusual and abnormal shopping behavior. For example, if someone puts an item from the store into their pocket, authorized store personnel will be alerted.

Real-time alerts
The open nature of a Video AI Analytics system allows for real-time alerts. These alerts are immediately passed on to authorized and trained store personnel, who can respond quickly.

The privacy of our customers is protected through privacy filters. This means that all faces are made unrecognizable and can only be viewed by authorized personnel if an incident has been reported. Footage from the smart cameras is not stored if no unusual or abnormal shopping behavior is detected. If such behavior is detected, the footage is generally kept for 28 days. After the test, we will decide whether we will implement this system in other stores as well. After the test, we will decide whether to implement this system in other stores.

For more information about the use of Video AI Analytics and the associated rights of our customers, we refer you to our Privacy Policy: SPAR | privacyverklaring van SPAR

SPAR Holding, Waalwijk June 202