Vision Intelligence: FLK-i®

Inspired by birds of prey.

See the bigger picture with FLK-i®. (The smaller picture too.)

There is a wealth of information to be mined out of every scene – be it one captured by a conventional video camera or a modern 3D LIDAR. Ignitarium’s FLK-i® platform applies AI-based Scene Analytics techniques to unearth those vital pieces of information that are important to you – be it for safety, compliance or automation.

Key Features of FLK-i

Accurate scene analytics on still and moving images

Support a variety of “image” sensors – RGB, IR, thermal, LIDAR, RADAR

Supports temporal and spatial analytics

Simultaneous Multi-sensor scene analytics support

Advanced Hybrid algorithms merging classical CV and DL techniques

In-field Customer Re-trainable architecture

Custom Neural Networks that minimize the requirement for vast training data sets

Multiple deployment modes – edge, in-premise server and cloud based

Solution scalable from low-cost micro-controllers to GPUs

FLK-i Product

TYQ-i Attributes

Uniqueness

The FLK-i platform is a library of software components that allows building a complete glass-to-glass imaging pipeline to perform complex scene analytics. Using advanced processing algorithms and proprietary neural networks, it is now possible to perform a mix of object, pattern and event analytics on input data streams originating from a multiplicity of diverse pixel and point cloud sensors.

In-field trainable SDK

SDK to allow quick deployment and re-tuning of pipeline for various application scenarios

Scalable across platforms

Rich library that enables porting of analytics components across diverse hardware platforms ranging from MCUs to GPUs

Minimal training set

Advanced custom neural networks that can be trained for high accuracy inference with minimal data sets

Case Studies

FLK-i in Action

Deployment Models

On prime server

On-premise server

Edge

Cloud Storage

Cloud

Attire & PPE Detection

Use cases :

  • PPE Checks
  • Disallowed attire checks

Use cases :

  • Non-intrusive adherence checks
  • Customizable attire checks
  • Post-deployment trainable

Behavior Monitoring

People Detected walking in a crowd

Use cases :

  • Fall Detection
  • Social Distancing

Highlights :

  • Can define region of interest to monitor
  • Multi-subject monitoring
  • Multi-camera monitoring
  • Alarm triggers

OCR / Pattern Recognition

Use cases :

  • Analog dial reading
  • Digital meter reading
  • Label recognition
  • Document OCR

Highlights :

  • Configurable for text or pattern recognition
  • Simultaneous Analog and Digital Dial reading
  • Lightweight implementation