State of the Art Technology

Our technology is based on state-of-the-art Artificial Intelligence, the latest in Computer Vision, sophisticated Computing Science and avant-garde Software Engineering.

Julia – futuristic software engineering, today

We use the programming language Julia to shorten the time from prototype to product. Julia combines the terse syntax of languages like Python and Matlab with the efficiency of languages like C. A program written in Julia requires less programmer effort, i.e. fewer lines of code, to yield a program of equal efficiency to a program written in C.

randmatstat written in C.
randmatstat written in Julia.

Julia is composable

Programs are composable; one program can be composed of other programs – like pieces in a jigsaw puzzle. Julia enables composability at a an unparalleled scale due to meta-programming features and the unique combination of multiple dispatch and Just-In-Time compilation. An example is the implementation of a machine learning library for artificial intelligence / neural networks using very few lines of code, by being composed of other Julia programming libraries, e.g. for GPU’s.

Julia is fast!

Variables in Julia have types. Types have made computer programs fast since the 1950’s and 1960’s. Surprisingly, types are absent from popular programming languages, Python and JavaScript.

Julia can execute across multiple processor threads – enabling efficient use of modern multi-core processors.

Julia can execute across the computation units of massively parallel processors like Graphics Processing Units (GPU’s).

IHP Systems has been applying Julia since 2014. Julia popularity has been growing rapidly since its launch in 2012. Source:

AI for Computer Vision

We employ and develop Artificial Intelligence (AI) systems for computer vision tasks such as object detection and recognition, e.g. for waste sorting robots – enabling the robots to perceive and act in a visual world.

IHP Systems have been doing machine learning systems for more than a decade; from feature engineering to current AI / Deep Learning systems based on Artificial Neural Networks.

Digital Codes for Identification and Tracking

We have developed digital coding schemes for identifying and tracking physical objects, e.g. parcels or packaging.

The codes can be written as digits and characters, but can also be printed as bar codes or QR codes, or embedded invisibly into a product design in the form of digital watermarks.

10 + 13 = 10 – 13 = 7

The codes are based on sophisticated computing science and mathematics, i.e. algebraic fields, where seemingly counter-intuitive logic like 10 + 13 = 10 – 13 = 7 protect the codes against counterfeiting and ensure reliable detection and recognition.

Clear Skies

We build on cloud computing components like Docker containers and Kubernetes, but provide all of our systems and services as cloud-free components to ensure privacy and integrity.

Open Source

Where ever possible, our systems are composed from open source building blocks – building on solid, reviewed software. We also contribute open source software to the rest of the world, including contributions to projects led by Google et al.

13 repositories, 0 followers.