Custom Development

Computer vision, deep learning algorithms and mathematical models

We have developed many computer vision applications in the fields of robotics, manufacturing security, advertising and education. It includes Volumetric Video Capture products, a market leading proprietary Depth Codec applications, robust point cloud and 3D mesh generating and filtering algorithms, object recognition and extraction and movement tracking.

Some of the projects we have delivered for our clients include:

– Depth Codec development: live point cloud streaming and compression

– Stereo vision and calibration software development

– Online 3D model player and editor

– Point Cloud manipulations, segmentation, filtering

– Mesh generation in real time and in post-processing

– Format converter development: PLY, OBJ, FBX, glTF, EVE, EFW, etc.

– Deep Neural Network models for manufacturing

– Point cloud filters and editing software.

– Depth sensors and Sensor fusion development

– Kinect and RealSense 3D viewer and editor

 

 

Integrating autonomous robotics system and scanning: AURORA – EU Commission project – Intelektas LT

The system consist of Volumetric Video system monitoring premises and enabling perceptual human-computer interface to control a mobile robot capable of autonomous navigation and mission based delivery. EF EVE web platform is facilitating the management and distribution of created volumetric content between doctors and operators of autonomous vehicle system and on-site VV capture. Learn more

 

Market Leading 3D Depth Codec

The EF EVE™ team has developed market leading proprietary EF EVE™ Depth Codec that solves main industry problems: compression, recording and live-streaming of Volumetric Video.
It has distinctively higher compression and faster algorithm than most of lossless and lossy compression solutions on the market. It has been applied in many industry cases where live 3D data streaming is needed
Get in touch to discuss your case.

 

Point Cloud filtering, segmentation and cleaning

We have worked on many projects in robotics, manufacturing and security where simple off the shelf sensors were used. These type of set ups usually generate noisy and inconsistent data. Therefore, we have developed bespoke functions to meet clients needs including:

  • Noise reduction
  • Meshing
  • Filtering
  • Segmentation
  • Surface smoothing

Intrinsic and extrinsic camera parameters calibration

Every project which involves 3D sensors requires camera calibration. Sensors have different number of parameters as for example, Intel Realsense has RGB and two IR sensors so there are three sets of intrinsic parameters while IR and depth intrinsic parameters match. Each of the parameters needs to be calibrated separately for each of the devices.

To automate this exhausting and time consuming process our team has developed solutions combining intrinsic and extrinsic camera parameters calibration.

 

 Contact us to discuss your case.

Start typing to see search results

Try "photogammery", "diorama", or "dinosaur", for some cool results.