In the realm of data-driven insights, FootfallCam consistently gathers market feedback, shedding light on critical areas of improvement for incremental product improvements.
Challenges: Device relocation without configuration updates, counter position adjustments, overhead obstructions, noise disturbances during operational and non-operational periods, and challenging environmental factors.
Recognising the significance of Intelligent Environmental Learning powered by AI, we address these issues to enhance accuracy and performance.
1.1 Auto Operating Hours Detection
Based on the conditions detected below, each device learns from historical trends using Machine Learning (Random Forest); can accurately determine the exact opening and closing time (minute resolution) of each day
(i) Traffic pattern
(ii) Lighting conditions
(iii) Shutter door status (if any)
1.1.1 Store Lighting On / Off
Environment lighting changes detection
1.1.2 Shutter Door Open / Close
Auto alert if the entrance is detected as close
1.2 Overhead Blockage Alert
Using 3D imaging, if there is an overhead decoration that is believed to have blocked the walk paths, an alert would be sent for further assistance.
1.2.1 Overhead Blockage Detections
Festive decoration(s) blocking the counter live view partially
1.3 Walk Path Changes Alert
When 'additional' objects were detected in the walk paths AND an abnormal traffic pattern is detected, the device would send out an alert for further assistance.
1.3.1 Change of Store Furniture
(i) Change of product display
(ii) Add / Remove furniture at the store front
1.4 Staff and Guard Exclusion
Counting the number of customers, not staff. This is particularly important for luxury stores where footfall are low.
1.4.1 5x Methods of Staff Exclusion
(i) Counting Line / Zone exclusion
(ii) Dwell Time exclusion
(iii) Wall Exclusion Button
(iv) AI Staff Tag
(v) Discreet Fabric Tag
1.4.2 Guard Exclusion
Based on the "hovering" walking path pattern; AI algorithm would classify the person as 'Guard'.
1.5 Change of Device Position Detection
Each device is continuously learning about its environment. If there is a global offshift of the environment, a "Device Change Position" alert would be sent out.
1.5.1 Device aware of its environment:
(i) Counter live view is showing different entrance environment
(ii) Might have reversed in / out direction
1.6 Device Relocation Detection
Customers often relocated the devices to another sites without updating the system. It caused invalid counting data. The new AI algorithm would detect an environment change and send an alert for further assistance.
Before: Device is paired with BranchID6213
After: Device remains with BranchID6213, however there is an abrupt environment change (detected by FootfallCam AI)
