Our Mission

Teaching Drones To See in 3D

Molfar Technologies’ researchers are developing high-efficiency semi-supervised machine learning algorithms and training data gathering automation methods for autonomous systems.



Our mission is to enable unmanned aerial vehicles to use their cameras to autonomously recognise and classify 3D objects and record their coordinates without constant operator supervision or data connection and navigate the GPS-denied environments with only UAV camera and inertial sensors

Find Out More

Subscribe

Sign up with your email address to receive news and updates

Why Us

Complete Autonomy

No need for high-bandwidth data connection in operating mode. Once trained, the UAV can operate autonomously, only alerting the operator when object of interest is found

High Reliability

Dynamic selection of optimal decision rules allows for high certainty 3D object recognition invariant to scale, rotation and angle – even when training dataset is small

GPS-denied navigation

Our technology allows navigation in GPS denied environments with only UAV camera and inertial sensors

Low Cost

Low computational cost and substantially reduced demand on hardware specifications as a result. Our algos run even on a Raspberry Pi

Our Approach

Adaptive Technology

At the heart of our proprietary technology lies our original research on self-organizing feature representation models with adaptive feature coding. This unique combination enables the system to select most cost- and information-optimal decision rules

Inspired by Nature

Our innovative use of nature-inspired search algorithms allows for low computational complexity in both training and decision-making modes. This makes our technology particularly suitable for autonomous device applications, where computational resources and amount of training data are constrained

Better Learning

Our advances in automation of training data collection substantially speed up the deployment and configuration of intellectual systems. Self-adaptation of multiple system parameters to optimize the information criteria makes the learning process speedier and more efficient