Updated: Aug 22
Watch Axis Communication's Stephen Joseph and Parabit's Rob Leiponis discuss the benefits of optimized video surveillance sensor placement, and how it can have a major impact on security and retail engagement.
Subscribe to our podcast on Apple or Spotify to listen to this podcast and receive updates on new episodes. Listen to the audio-only version with playback controls here.
Hilary Kennedy: Hey there, welcome to A Bit About, a Parabit Systems podcast. I'm your host, Hilary Kennedy, and today's episode is going to dive into how technology is enabling a smarter and safer world by creating network solutions that provide insights for improving security and some new ways of doing business.
And I have two great guests joining me on the show today to share their insight. First, Rob Leiponis, President and CEO of Parabit Systems, which is a leading global provider of innovative hardware and software solutions serving critical infrastructure. And since founding the Long Island based company in 1995, Rob has driven the development of technology deployed by hundreds of financial institutions throughout North America and Europe. And Rob draws on tremendous experience that spans 35 plus years.
And then our second guest is Stephen Joseph, Segment Development Manager of Banking and Finance of Axis Communications. And Steven is responsible for strategy development for the banking segment, education development for internal and external customers. He speaks at national conferences and has a host of other critical roles. So it goes without saying that both of them will have a wealth of knowledge to share on improving security, and how cameras are enabling data analytics that the world needs.
So welcome to the show, gentlemen.
Stephen Joseph: Hi Hilary, thank you.
Rob Leiponis: Thank you.
Hilary Kennedy: Alright, so first, I want to start off with effective facial image capture. It supports safety and security, as well as you know, retail and marketing. So what are some of the challenges, though, when it comes to capturing good facial images?
Stephen Joseph: Well, for the most part, capturing good facial images are really important, especially when you're dealing with customer-facing environments. Specifically, and for me, it's most of the time in the banking and finance institutions, and it can also be in retail as well, where you see customers coming on a regular basis. And one of the main challenges that a lot of these businesses have is being able to place cameras at an optimal location to be able to capture great facial images.
When incidents occur, and law enforcement is called in and they need to capture or be able to get evidence as it relates to who may have perpetrated a crime, being able to have that optimal image to be able to recognize and provide forensic detail is really, really important in those use cases.
Hilary Kennedy: And, Rob, do you have any thoughts on that too? You know, what are some of those challenges when it comes to capturing good facial images.
Rob Leiponis: Just basically eliminating any type of obstructions that can come from any existing cameras that they would like to leverage for analytics, to installing or repositioning a more discreet camera that's closer to the actual face to eliminate the possibility of someone coming in obstruction, of all the images of the people's faces as they walk in and out of retail facilities, transportation hubs, you know, loading docks and logistics facilities.
I think it's important, from an access control perspective, from an artificial intelligence perspective, to be able to leverage that technology that's so great, and that's been helping create such a life-safety environment for the world. To be able to maximize the investment by placing those images in the sensors close to people's faces.
Hilary Kennedy: It's so true. It's helped in so many countless different situations, especially where safety's involved. And I heard both of you mentioned analytics and so Steven, I want to direct this at you can you give an example of the capabilities of analytics when effective facial image capture is achieved?
Stephen Joseph: Absolutely. There's a lot to know with regards to facial capture. People use the term interchangeably, but there are different levels of capturing faces. You have traditional face capture where you just you know, capturing an image of a person's face, you know that it's an individual, whether it's male or female. You have facial detection, where you're able to actually detect that it's a face. Facial recognition, being able to recognize who that individual is. Facial identification, where you're actually able to, with 100% certainty, be able to determine that that individual matches face that you currently have in some type of database.
And you see this a lot in law enforcement where they're looking through databases, and we see this a lot on TV where you see somebody looking through a mug shot book, right? They're trying to do forensic identification. And they're comparing faces that are known faces, to help someone identify that person.
And when you have proper placement of cameras or video cameras in these cases, and if you walk into most establishments, today, you're going to see a camera. It's either going to be on the ceiling, or it's going to be on a wall. Normally, it's far back from where that person is actually standing. But the types of topics that we're trying to discuss today has to do with having a proper placement of an image sensor that allows you to get a great shot of a person's face.
When it comes down to facial detection, you need about four pixels. And you think about pixels and you think about let's say 1080P or 4k television, there are pixels in that image. You need about four pixels across a person's face to actually get facial detection. In order to get facial recognition you need about 20 pixels across the face. To do identification, you need about 40 to 80 pixels across the face. So that gives you an example of the different levels of facial capture, facial detection, that you can look be looking for.
Hilary Kennedy: Well it's so interesting, because you can't go in any store anymore without you know being able to see the cameras. So Rob, I would love for you to tell us a little bit about when it comes to retail and marketing. How is facial image capture the most beneficial there?
Rob Leiponis: So, facial capture is good for determining what type of content that you may want to display on various digital signage, or sending messages to tellers in a platform that a certain person with a certain profile has walked in, or back they may be a customer of yours. So that way you can target or create more targeted questions in order to convert that customer to a better service that they may be providing for them.
Notification of a customer arrival like in a Nordstrom experience where you know when a customer arrives, there's a sales rep that's assigned to that person every time they walk into the store. Through analytics, that person could get notified immediately that their premier customers are walking in the southeast entrance of the building and provide a better greeting experience for them.