security solutions
Flux Tensor helps surveillance operations and analysts focus on pre-screened objects and events of interest. pFlux deemphasizes nuisance noise and irrelevant object movement to cue attention to critical events and object behaviors. With over 1 billion security cameras currently deployed worldwide, technologies able to make sense of this massive data stream are critical.
The pFlux Security Solutions perform effectively regardless of:
- Image/video resolution
- Video type (RGB, IR, etc.)
- Nuisance motion caused by weather, occlusions, glare, reflection, camera movement
Persistent Change Detection
The videos below demonstrate a math-based approach to differentiating a body in motion from an object coming to rest. The upper right quadrants show a color-coded output of the pFlux Solution used to draw attention to an object-left-behind scenario (red pixels). This was achieved in near real-time (~250 ms) without deep-learning or prior AI/ML modeling or training. The solution exhibits consistent performance in all camera views/scenes regardless of time of day, shadowing or background noise. The examples below show the object detection and tracking to be perform equally well in day and night conditions.
Persistent Change In Day
The upper right quadrant shows the pFlux Solution output, highlighting a body in motion in blue while the bag-like object exhibiting persistent change is highlighted in red. The solution identifies the persistent change behavior equally well at night and when the object is in the distance, nearly undetectable to the human eye.