Computer Vision
Smart advertising
Smart advertising

Thanks to its Cognitive video analytics and computer vision solutions, EZD developed many solutions to convert “plain old advertising” in “Smart advertising”, ere some:
Usually, when a viewer looks at an ad, they may only see the message for a few moments before looking away. However, with an interactive ad that responds to the viewer’s gaze, they’re more likely to look longer to see what will happen. Thus, views last longer, and the message has more time to sink in. (It’s important to note that video ads for the sake of movement is not what we’re talking about here. The movement needs to be a part of the message to really be effective at enhancing the overall ad.)
Taking the concept a step further, imagine an ad that ‘talks’ to the viewer. Since the technology gives ads self-awareness, a donut shop could create an ad that says good morning to anyone that walks by, or a clothing store could create an ad that compliments (or mocks!) outfits in the crowd. The technology would also allow advertisers to incorporate a video that starts only when someone is looking, rather than playing over and over again on a constant loop.
In addition to enhancing the message, an outdoor ad that’s aware of when people are looking at it ushers in a whole new level of measurement, as view numbers no longer need to be rough estimations of foot traffic and awareness. Instead, each ad can be bought and sold based on accurate view numbers and actual engagement, giving advertisers proof that they’re getting what they’re paying for, and allowing media companies to price their high profile ad placements with the premium they deserve. That’s why we call it Smart advertising.
Digital Signage Audience Measurement
Digital Signage (also known as Digital Out Of Home, or DOOH) is among the fastest growing media these days. This is hardly a surprise: marketers are enticed by the ability to finely define broadcasting strategies down to every single screen, by the ease with which content is updated and by and the proximity to the purchasing act. Nonetheless, most digital signage networks still lack the solid audience metrics necessary to prove their impact to the whole value chain of stakeholders, from location to advertisers and agencies.
Our solution has been designed to address this very “hot” need. Running on the same architecture as the video player, it analyzes the video stream provided by a standard low-cost webcam or IP camera to offer a rich variety of audience metrics 24/7 on every measured screen: actionable data include the number of viewers, the Opportunities To See, attention and dwell time figures, as well as the audience’s demographic distribution (gender and age group). These anonymous audience metrics are uploaded to our center, where they are aggregated into meaningful charts and tables; they can also be correlated with playlogs to deliver audience metrics by media – a level of precision never seen before!
You can use our solution to
- Assess the true impact of your whole Digital Signage network, by site, by type of location… down to any single screen at any given time.
- Analyze your audience year-long and discover trends in time and differences among demographic profiles
- Provide solid data to your advertising prospects to demonstrate your reach
- Provide a rich audience report to each of your advertisers at the end of their campaigns.
Retail Metrics
The triggers which lead to purchases in retail are still largely unknown. Lately, a number of industry initiatives and projects from large FMCG companies have focused on possible frameworks to understand how shopper behavior in stores; how customers peruse specific categories; why they eventually buy; and what is the variability across people. Quividi offers retailers, brands and research agencies an elegant solution to finely analyze shopper behavior in key retail zones.
Small and unobtrusive video sensors are placed along the shelves, either at regular intervals, or in specific zones, or embedded in a branded fixture; the sensors provide multiple video streams that are analyzed in real time by VidiReports running on a nearby PC. VidiReports counts and classifies face exposures determining ow many shoppers looked at a zone on a shelf and for how long. Fully respecting visitors’ privacy, VidiReports does not record any image, nor does it record any uniquely identifiable data.
Adaptive Marketing
The Right Message to the Right Audience, in Real Time
The Internet has quickly gained a tremendous share of the advertising budgets thanks to its superiority in terms of precise targeting and accountability: online media sellers can deliver a measurable performance toward specific groups and bill campaign “per click”. Live face detection solutions can bring similar benefits to any organization communicating on Digital Out of Home media, while fully respecting viewers’ privacy and legislation.
Our solution analyzes the video stream produced by a small sensor placed nearby a screen and reacts in as little as 1/10th of a second to produce a live description of the audience. This description includes the position and distance of all detected persons, their attention state (i.e., are their faces turned toward the sensor), as well as their gender. (Faithful to its privacy pledge, we don‘t save any image or any uniquely identifiable data.)
Third party applications can easily adapt their programming by listening to our communication protocol and setting their playlist strategies to put forth the right message to the right audience. we already offer ready-to-use integration schemes with the leading digital signage software suites on the market.
You can use our solution to:
- Display specific messages depending on the appearance of a certain demographic profile or a given viewer’s behavior or position
- Engage viewers in fun activities (e.g. get them to spin a virtual 3D object by just moving their head);
- Activate the sound only when customers are watching the screen to reduce employee fatigue;
- Quickly identify the best performing messages from a pool for a particular target group;
- Bill advertising campaigns by actual impression, and offer advertisers a guaranteed number of impression that fits their precise budget.
Other technologies and applications may include:
Dispensing a fragrance depending on the position and profile of the shopper (see the specific case study);
Adjusting the sound animation in a store according to the dominant shopper profile at any given moment.
Small laser projectors to project a touch-sensitive display everywhere.
Photo Gallery



