Merging Real & Synthetic Worlds for Computer Graphics and AI

We create realistic training data for your AI application in computer vision.

Why rabbitAI? Newsletter
Description Description

Showcase

A Small Glimpse at What We Do

Stunning 3D Reconstruction
Very detailed depth reconstruction. We create a high-precision depth map in one shot for each image. No motion blur, real environments and a depth value for each pixel (dense depth). We can enhance your Augmented Reality (AR) AI with a mm accurate depth map. Relevant for Virtual Reality (VR) and Augmented Reality (AR).
Ultra Realistic Augmentation for Rare Edge Cases/Corner Cases
We add ultra-realistic objects in real scenes. Features include correct lighting, occlusions, reflective surfaces and different weather situations. Augmentation is taken to a new level, that's why we call it Mixed Reality (MR).
Request a demo!

Why rabbitAI?

Towards Better AI

Ensure Safety and Create Trust Autonomous systems like autonomous vehicles need to react reliable in close to 100% of the cases. We create accessible and independent testing standards for such systems based on computer vision. We can add variability to any AI testing data set with ultra realistic and relevant corner cases to boost your neuronal network.

Reduce Development Cycles Autonomous systems are a complex interplay of hardware sensors and acting AI. System designers face a chicken-and-egg problem: Training data needs to be acquired with hardware for which optimal design can only be derived after AI has been trained. We break this dependency to shorten development cycles. Our training data can be adapted to any existing camera system.

Our Offers

How We Help Your Business.

Fast and Accurate Captures of the World Our unique recording technology captures complex and dynamic scenes into precise and feature-rich digital models. With only a single shot. This way we get very precise depth reconstruction for every pixel in the scene without motion blur.

Creating New Worlds A polar bear in your backyard, mirrored in a puddle, at night! Sounds crazy? Just a click away with our technology: We enrich your training data by adding relevant objects. This approach goes beyond Augmentation, that’s why we call it Mixed Reality (MR). Mix and match our captures, add your own models to achieve ultra realistic augmented scenes with correct lighting/illumination and physics. We create the corner cases that matter to you.

GDPR Compliance All our services are GDPR compliant. The mix-and-match approach avoids capturing personal data at it’s core.

Evaluation Need to choose sensors for your AI application? Got stuck with your AI performance? Our data driven world models allow comparing different sensor setups in advance and help to systematically identify your data bottlenecks.

Get in touch!

How Does It Work?

Acquisition, Augmentation and Quality Assurance. All in One Go.

Capturing Real Scenes

Real-world data captured with our own recording techniques guarantees optimal data sets including precise 3D depth reconstruction/maps.

Creating Safe Worlds

Increasing data richness and variability by creating new scenes out of already captured ones. These can be adapted to your camera setup.

Humans in the Loop

We provide the highest quality by always having humans in the loop. At the same time we reduce QA efforts by reusing high-quality material.

Our Team

Hendrik Schilling

Co-founder | CTO

Dr. Marcel Gutsche

Co-founder | COO

Karsten Krispin

Co-founder | CEO

Dr. Clemens Velte

CBDO

Prof. Dr. Carsten Rother

Mentor