Stage 01
Capture
Take photos of the object with any camera. This is the only manual step.
Computer vision & machine learning R&D · Est. 2011
We teach machines to see.
Aurvis is a research-driven company in Ankara building computer vision software, from photogrammetric 3D reconstruction to star trackers for satellites.
ScrollTrusted by research institutes, enterprises and museums
Solutions
Research that ships as products.
Over a decade of computer vision research, turned into tools for 3D modeling, space, accessibility and culture.
013D Reconstruction
Pixel Scanner
Hundreds of photos in, calibrated cameras, dense point clouds and textured meshes out.
02Space
Star Tracker
Lost-in-space star identification, robust to radiation and false stars.
03Accessibility
Low Vision Assistant
Real-time, high-contrast board view with the teacher removed.
04Mobile · Recognition
Smart Museum Guide
3D object recognition turns any smartphone into a rich museum guide.
05Space · Ongoing research
Lunar Navigation
Recognizing craters to tell a spacecraft where it is in lunar orbit. Ongoing with TÜBİTAK UZAY.
Pixel Scanner pipeline
From photographs to 3D models in minutes.
Hundreds of images in. Calibrated cameras, dense point clouds and textured meshes out, with no manual work in between.
Stage 02
Calibrate
Camera intrinsics and poses are recovered from the images alone.
Stage 03
Densify
A dense, oriented, colored point cloud is computed.
Stage 04
Mesh
A triangulated, textured mesh is extracted, ready to use.
About us
A small, research-driven team with PhDs in computer vision, building software that understands images.
Aurvis provides products and consultancy in computer vision, machine learning, image and video processing, visualization and high performance computing. We were founded in 2011 with the support of the Turkish Ministry of Science, Industry and Technology, and spent our first two years hosted at the Microsoft Innovation Center in Ankara with Microsoft's support.
2011Founded
15+Years of R&D
8Reference partners
Team
Small team. Deep expertise.
Strongly research oriented, with backgrounds at EPFL, Microsoft Research and MERL.
Engin Tola, PhD
Founder & CEOPhD in Computer Vision from EPFL. MSc and BSc in Electrical & Electronics Engineering from METU. Previously at Microsoft Research (USA) and the EPFL Computer Vision Laboratory.
Geometric vision3D reconstructionCalibrationReal-time
Zafer Arıcan, PhD
Senior AssociatePhD in Computer Vision from EPFL. Degrees in Electrical & Electronics and Computer Engineering from METU. Previously at Türk Telekom Ar-Ge, Argela, Comodo and MERL.
Augmented realityObject recognitionDeep learning
Görkem Demirtaş
Senior Software EngineerMSc and BSc in Computer Engineering from METU. Previously at SpeedDate, Invento and the METU Intelligent Systems Lab.
MobileGPU programmingPython / C++Cloud
Research notes
From the lab.
Ongoing · with TÜBİTAK UZAY
Crater-Based Lunar Navigation
Where is the spacecraft? Detecting and identifying lunar craters to recover a camera's position and attitude in orbit.
Read note →
26 Aug 2017
Image to LaserScan Registration
A new method for calibrating RGB images to laser-scanned environments, developed for Autodesk.
Read note →
26 Apr 2017
Star Identification
How a lost-in-space star tracker recognizes constellations even when radiation adds many false stars.
Read note →
Start a project
Have a vision problem?
From research prototypes to production software: tell us what you need machines to see.
Mustafa Kemal Mh. Dumlupınar Blv. ODTÜ Teknokent MET Yerleşkesi, No 280 B-Blok No 14, Çankaya / Ankara, Türkiye