Panos Moutafis, Ph.D. is Co-Founder and President of Zenus Inc., which specializes in facial recognition systems for multiple applications. A computer scientist by trade, he is well-known for his work ethic, diligence, and persistence. I caught up with him over breakfast in Houston recently, keen to discover if facial recognition technology could become mainstream, or sit on the fringes like many other great ideas such as RFID and iBeacons.
Facial Recognition: Fact or Fad?
Peter: Panos, first, is facial recognition technology in its heyday?
Panos: This is more than a cool technology that people will start using ‘sometime in the future’. Facial recognition adds clear value such as improved security and better user experience along with unmatched analytics and insights. It is already used to tag photos on social media, unlock and pay with cell phones, and go through border security. Since people are getting accustomed to the technology, facial recognition is becoming an expected service rather than a nice to have feature. 2018 is the year innovative events are embracing facial recognition!
Peter: And this goes beyond what most people imagine?
Panos: Couple it with adjacent technologies such as emotion recognition or human posture recognition and the possibilities are limitless. You gain full control. Every powerful technology comes with great responsibility though. Respecting people’s privacy and handling metadata in a transparent and prudent manner is vital. Every party who has access to this type of information must be thoroughly vetted, 100% trustworthy and held accountable to the highest standards.
Event Marketing: Let’s Get Personal
Peter: Let’s talk about events. How does this technology work here?
Panos: The use of physical tokens and privileged information have become integral parts of an event lifetime. They are used to check in people, restrict access, personalize the experience, measure attendance, extract analytics, and perform lead retrieval.
Facial recognition belongs in the third form of authentication along with other biometric approaches. It is a software, which can identify a person from a database of faces without requiring a physical token or the user to provide any privileged information. Technological advancements have increased accuracy and drastically reduced. Therefore, we are seeing increased adoption in other industries (e.g., airports, social media, and cell phones).
Peter: Well, take us from the start of the process; I always chuckle to myself when I check into to a high tech event for a mega-company and I wait in line for someone to tick off a spreadsheet or hand write my badge!
Panos: The attendees are not always good at following instructions displayed on the terminal. They often cannot find their ticket; it takes them a while to retrieve the QR code on their phone, or they simply need time to type in their email address.
Facial recognition addresses the root cause of the problem by eradicating the need for user actions. Depending on your preference, you may want to implement a self-service or a hosted check-in mode. In either case, the premise is the same. When an attendee approaches the check-in station they will be instantly identified without them having to take any action. Simple.
The best part is that you do not need to purchase expensive hardware because any device with a camera works. This includes virtually all laptops, tablets, and smartphones. In addition to making the check-in process more efficient and thus reducing costs, face recognition has been proven to increase user engagement and attendee satisfaction. It is an excellent way to promote the innovative nature of your brand and impress attendees.
Peter: And I guess ID checking is enhanced?
Panos: Even though checking the ID of the attendees entering the venue is a step in the right direction, most of the existing implementations have limitations. Attendees are asked to present their ID and an untrained host is taking a quick look before giving it back. This is not enough. Doing a proper ID check, on the other hand, is time-consuming. One would have to check the issue and expiration dates, scan the code and make sure it matches the ID unique number, check whether the picture matches the person presenting the document, and so forth.
There is a faster and more secure way to perform this task properly. In particular, attendees can be requested to take a picture of their ID along with a selfie when they register online. A facial recognition algorithm will ensure that it is a real picture (liveness/spoof detection) and that the two faces match; other computer vision algorithms will perform the rest of the checks in real-time. We see this approach being used widely in other applications such as the banking and hospitality industries (e.g., Airbnb).
Peter: Could this be used for Session Tracking as well?
Panos: Event hosts spend a significant amount of time recruiting speakers, curating the content of their shows, and organizing sessions that cover different aspects of the event theme. In addition, conferences with a deep educational focus often assign credits and accreditation to participants who attend specific sessions.
Face recognition is a good fit for this because it requires minimal setup and extracts analytics in a non-intrusive manner. Depending on different technical factors, session tracking with face recognition can be as simple as putting a tablet or cell phone on a stand near the entrance of each room. The camera will automatically capture the video stream and send it to the cloud for processing. There is no need for special hardware and expensive installation costs.
Peter: We’ve experimented with heatmaps for different clients using RFID technology in the past – could this be done more cost-effectively now?
Panos: Face recognition can be a great way to compute and draw heatmaps. Tracking the number of faces visible by the camera across the different event locations is straightforward. It does not require special hardware and it is easy to install and configure. Depending on the level of investment, the information can be as high level as how many people on average stood and passed by a certain point of interest or as fine-grained as extracting insights by group type and knowing each attendee’s journey.
Peter: How about other applications out on the show floor?
Panos: One can use screens (equipped with a camera) around the venue that identifies attendees and display personalized information ranging from where food is being served to what next session they should attend. Likewise, the information desk personnel could be equipped with similar capabilities. Being able to identify a person while they are walking towards the desk allows them to personally greet them and anticipate their needs.
Along the same lines, one could combine face recognition with chatbots to offer the most efficient and personalized customer service. There is a myriad of interactive applications that could be developed. It is up to your imagination.
Calling All Marketing Qualified Leads
Peter: And my favorite subject of course: Lead Retrieval! How does face recognition factor here?
Panos: As you know, the current situation is that exhibitors have to manually scan each person’s badge and enter all the relevant information into a lead retrieval application to summarize the discussion. With face recognition, it is possible to automatically keep track of how many and which people visited the booth. The lead generation and conversion statistics are vital in the sales process.
In addition, this technology allows you to keep track of how long attendees stayed at the booth and how they were feeling. Analyzing their sentiments combined with the information collected before/during the show enables lead management like never before.
At the end of each day, the exhibitors will receive a full report. The leads will be automatically scored by their likelihood to make a purchase decision so that the team can focus on the best targets and following up promptly. Face recognition combined with audiovisual sentiment analysis will have a dramatic impact on leads retrieval.
Peter: I know that some people will think that Facial Recognition is creepy and they feel uncomfortable about it?
Panos: That is true, but real-world deployments by Zenus and our partners show that the majority of attendees are willing to use facial recognition. That is, the first time an event planner introduces the option they can expect 50-60% of the people to upload their photo. We see attendees talking to their friends and colleagues about how fast and easy the check-in process was, so we expect the opt-in rate to grow for repeat events.
The great thing about facial recognition is that not everyone must opt-in. Instead, it works in conjunction with name search and barcode scanning. The same check-in station can simultaneously allow attendees to use any of these three ways to get their badge. If you have enrolled in face recognition you will simply go through faster!
It is worth noting that a participation rate of 20-40% is enough to make a difference to the on-site registration process. Speeding up the process for this portion of the registered attendees can address bottlenecks.
Peter: What about collecting the images – is that hard?
Panos: There are plenty of online registration companies that already allow attendees to upload their headshot or take a selfie. Most devices such as laptops, tablets, and cell phones have a front-facing camera. If necessary, one could opt to use their social media profile picture. It really isn’t as hard as people might think.
Peter: OK, let’s talk money – it still sounds expensive?
Panos: Some technologies are expensive, especially the ones that require specialized hardware, customizations, and on-site support. However, software-based applications tend to be more affordable. Luckily, face recognition falls into this category. Unless the event planner has excessive requirements, the associated investment is just a few cents per attendee expected to register.