THE KINETIX ADVANTAGE
We partner with companies to become their data analytics team integrating the kinetix platform modules allowing rapid development and deployment of intelligent products.
STRATEGY AND DESIGN
By engaging early in the process we can provide insight for potential new features and map out an in depth strategy and take it through to full design.
DATA
ACQUISITION
Outline data acquisition scope requirements and work with partner companies providing tools and knowhow to gather optimal datasets.
ANALYSIS AND OPTIMISATION
Based on hardware and power constraints we analyze multiple solution modules to find the best fit to map data to output features.
INTEGRATION AND DEPLOYMENT
Once offline optimisation is complete we integrate the solution to your products embedded chipset.
MOTION ANALYSIS
BEHAVIOUR DETECTION
INDUSTRIAL IOT
AUDIO DETECTION
MOTION RECOGNITION WITH KINETIX
Data Feature Engineering
The initial phase gathers real world datasets and create metadata tags that cleanly identify data segments. Context dependent features are extracted through data processing before classification.
Train Classification modules
Preprocessed data is tested against multiple classification modules and parameter sets. Classifiers include random forests, regression trees and neural nets. Comparative analysis finds the optimal module given the hardware and context requirements.
Test and Validate
Trained modules are integrated into firmware and tested in context where ground truth is known. Solution parameters are tweaked until required accuracy and sensitivity are achieved for the project.
KINETIX ANALYSIS TOOLS
VISUALISATION AND ANALYSIS TOOLS
In house custom toolsets allow visualisation and analysis of datasets and feature outputs. Ability to fine tune for speed and accuracy. Allows system testing before deployment to IoT devices.
