To Top

Welcome
to DIGIPRO

Objective

We will create 3D models

The objective of the DigiPro project is to develop a portable 3D scanner system comprising of hardware and software components that provides photo-realistic 3D reconstructions of rigid objects. The procedure should be fully automatic and output a smooth, closed mesh with associated material properties, which can be interactively visualized in a web browser for applications such as e-commerce or virtual museums.

The overall system includes three components: Acquisition Hardware, Reconstruction Software, Web Based Rendering. The whole process is fully automatic and provides a final photo-realistic 3D model in a few minutes.

Acquisition Hardware

The acquisition hardware consists of a projector and several cameras together with a turntable, where the object is being placed and thus able to be captured from various directions. An additional arm equipped with multiple LED light sources can be rotated independently around the object to illuminate it from different directions.

Reconstruction Software

The hardware is bundled with a fully integrated 3D reconstruction software that computes a digital representation (shape and appearance properties) of the object based on the acquired images.

Web Based Rendering

The resulting object representation is compressed by an optimization module to allow for interactive 3D visualization (change of viewing and illumination conditions) in web based applications such as online shops or virtual museums.

How We Will Work

The implementation plan consists of seven work packages.

Specification

Development of detailed specifications for the overall system.

Acquisition device

Development of the acquisition hardware including a software for controlling the device.

Reconstruction software

Development and integration of the reconstruction pipeline.

Data optimization and visualization

3D model optimization for web based rendering.

Integration and testing

Integration of the components, testing and optimization.

Project Management

To coordinate R&D efforts and apply risk management, to organise IPR issues and commercialisation arrangements.

Dissemination

Involvement of end users, website, results communication.

Partners

Who We Are

Our success depends on the strength of our team. Each partner has distinct expertise in the respective field of one the 3 defined components (hardware, reconstruction software, web based rendering).

AG Augmented Vision / Technische Universität Kaiserslautern

The department Augmented Vision of the University of Kaiserslautern has a strong background in the field of photo-realistic 3D reconstruction. It was heavily involved in all stages of the development of OrcaM, the world's first full-spherical device for simultaneous geometry and reflectance acquisition. Besides OrcaM, the department Augmented Vision has developed several other geometry scanners.

AG Augmented Vision / Technische Universität Kaiserslautern
http://av.dfki.de/

SPbSU Research & Development Center of Information Technologies Limited Liability Company

Having developed an own codesign language HASCOL, SPBSUIT has extensive experience in the design and development of efficient and cost effective hardware. It has a track record in the development of turn key industrial solutions integrating hardware and software and providing user friendly interfaces. It also has experience in computer vision systems, developing several applications for image detection and gesture control.

SPbSU Research & Development Center of Information Technologies Limited Liability Company
http://itc.spbu.ru/Profile.html

SilkTech Computer Informatics Solutions S.A.

SilkTech is a company that specializes in web applications and the creation of web sites / e-shops. SilkTech is also actively engaged in research projects related to web standards, digitization, ontologies, web optimization and even 3D web based environments.

SilkTech Computer Informatics Solutions S.A.
http://www.silktech.gr

Department of Electrical and Computer Engineering / Aristotle University of Thessaloniki

Algorithms and architectures for high performance computing signal processing and transformations. The AUTH team will be responsible for the mathematical modeling, design and analysis of three dimensional high order filtering procedures and their implementation on parallel multicore and portable GPU processors.

Department of Electrical and Computer Engineering / Aristotle University of Thessaloniki
http://www.auth.gr/ee


We want to change
the e-commerce reality

3D reconstruction offers a way for the automatic generation of product presentations that are realistic and can be explored by the user interactively. This enhances the shopping experience, while at the same time enables a better selection process for the customer.

We Are Programmers
We Are Coders
We Are Developers
We Are Creative
We Are Designers

— Welcome to the 3D enhanced e-commerce experience —

The project

A photorealistic and cost effective scanner for automatic 3D reconstruction

The objective of the project is to develop a beyond state of the art solution for photorealistic and cost effective 3D object digitalization, addressing an identified market need arising from the evolution of ecommerce. The procedure will output a smooth, closed mesh with associated material properties, which can be interactively visualised in a web browser. The overall system includes three components: (1) hardware consisting of a turntable, where the object is being placed, a projector and several cameras that capture pictures from various angles, (2) a 3D reconstruction software and (3) an optimisation module for use of 3D models on the web and mobile. The whole process is fully automatic and provides a final photorealistic 3D model in a few minutes.

The proposed solution of a cost effective and portable 3D scanner with photorealistic features enables to address the mass market of e-commerce, since it will allow the easy creation of 3D models for web stores. The target market consists of: (1) larger companies selling products over the web and (2) service providers that will offer digitalisation services to smaller companies.

Background

State of the art in this field

The development will be based on existing state of the art technological platforms developed by the partners. The Research group Augmented Vision of the University of Kaiserslautern has developed the Orcam system, which is designed to reconstruct objects in a single pass. The objects are placed inside a large sphere and digitized fully automatically. The scanning is based on the structured light principle; the diffuse texture and reflectance map is computed using images of many different lighting situation, created with help of the 600 LEDs. The resulting point clouds is highly précised, but of a size of several Gigabytes. Through appropriate meshing and maps creation, an effective object representation of 5 to 10 Megabytes is generated. The 3D visualization occurs in standard 3D viewers, using classical OBJ or Collada formats.

Economic
impact

Economic impact and added value of the proposed consortium

Realistic 3D reconstruction with cost effective systems is an actual market need arising out of new generation e-commerce. E-commerce is currently exploding. The most important factor in e-commerce is detailed and informative product description, as described in a report by ECC Handel. Interactive realistic 3D product presentation is the ultimate form of rich user experience and it becomes a must especially for luxury products. This requirement is even more important in combination with customization and personalization that represents one of the major trends in fashion. Realistic 3D models will both increase the conversion rate and decrease the return rate.

Societal
impact

Societal impact and added value of the proposed consortium

The possibility of creating precise 3D models in a cost effective way will allow also smaller providers to benefit more from e-commerce, supporting access to remote markets.
Added value of the proposed consortium and its complementary innovation potential: The combination of the complementary technological resources of the partners enables the fast track implementation of an automatic geometry and colour scanning system for 3D product digitalization for rich online and in-shop user experience.

Scientific
objective

Scientific and/or technological objectives

The project DigiPro will yield a solution for photo-realistic and cost effective 3D object digitization beyond the state-of-the-art. The envisaged system will consist of a portable scanning hardware and a reconstruction software, which will output a smooth, closed mesh with associated material properties. Optimization of the data complexity will allow to interactively visualize the reconstructions in web browsers by e.g. embedding them into an online web shop. By this technology, we address an identified market need arising from the evolution of e-commerce.

Our Work

Some Recent Objects

Web rendering is performed by filemonger.com, an online web viewer for design and 3D files by Vasileios Gougoulidis of Programmable Bit, Inc. The service is available from anywhere you have access to a browser or from your iPad/iPhone and Android devices. This service is kindly provided pro bono for the use of DigiPro.


Copyright

Statues of Apoll and Vesta were kindly offered from the Prussian Palaces and Gardens Foundation Berlin-Brandenburg (Stiftung Preussische Schlösser und Gärten Berlin Brandenburg).
Stiftung Preussische Schlösser und Gärten Berlin Brandenburg

For the data sets of Lehmbruck's "Female Torso" and Scherer's "Little Girl" we kindly thank the Museum Pfalzgalerie Kaiserslautern.

PUBLICATIONS

TU Kaiserslautern Project Publications
  1. Fetzer, T., Reis, G., Stricker, D., Working Title: From Correspondences to Point Cloud: Robust Calibration for Structured Light using two Cameras and a Projector, To be submitted to 3DV, (2018)
  2. Noll, T. . Efficient Representation of Captured Geometry and Reflectance, PhD Thesis, (2015)
  3. Köhler, J., Enhanced Usability and Applicability for Simultaneous Geometry and eflectance Acquisition, PhD Thesis, (2015)
  4. Köhler, J., Nöll, T., Schmitz, N., Krolla, B., & Stricker, D. . Structure from Motion in the Context of Active Scanning. Proceedings of the 10th International Conference on Computer Vision Theory and Applications . International Conference on Computer Vision Theory and Applications (VISAPP-15), 10th, March 11-14, Berlin, Germany, 620–628, (2015).
  5. Nöll, T., Köhler, J., & Stricker, D. . Robust and accurate non-parametric estimation of reflectance using basis decomposition and correction functions. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 8690 LNCS(PART 2), 376–391, (2014).
  6. Köhler, J., Peters, J. C., Nöll, T., & Stricker, D. . Improvement of Phase Unwrapping Algorithms by Epipolar Constraints. Proceedings of the 10th International Conference on Computer Vision Theory and Applications. International Conference on Computer Vision Theory and Applications (VISAPP-15), 10th, March 11-14, Berlin, Germany, 472–479, (2015).
  7. Nöll, T., Köhler, J., Reis, G., & Stricker, D. . Fully Automatic, Omnidirectional Acquisition of Geometry and Appearance in the Context of Cultural Heritage Preservation. Journal on Computing and Cultural Heritage, 8(1), 1–28, (2015).
  8. Gava, C. C., & Stricker, D.. A generalized structure from motion framework for central projection cameras. Communications in Computer and Information Science, 598, 256–273, (2016).
  9. Krolla, B. . Heterogeneous Reconstruction Approaches for Object and Scene Representation, PhD Thesis
  10. Ansari, M. D., Golyanik, V., & Stricker, D. . Scalable Dense Monocular Surface Reconstruction. Proceedings - 2017 International Conference on 3D Vision, 3DV 2017, 78–87,(2018).
  11. Sarkar, K., Varanasi, K., & Stricker, D. Learning quadrangulated patches for 3d shape parameterization and completion. Proceedings - 2017 International Conference on 3D Vision, 3DV 2017, 383–392,(2018).
  12. Sarkar, K., Varanasi, K., & Stricker, D. 3D shape processing by convolutional denoising autoencoders on local patches. Proceedings - 2018 IEEE Winter Conference on Applications of Computer Vision, WACV 2018, 2018–January, 1925–1934,(2018).
AUTH Project Publications
  1. Dimitris Floros, Tiancheng Liu, Nikos Pitsianis and Xiaobai Sun. Sparse Dual of the Density Peaks Algorithm for Cluster Analysis of High-Dimensional Data.Proc. 2018 IEEE High Performance Extreme Computing Conference (HPEC), 2018. [pdf] [Supplementary material] [BibTex]
  2. Iliopoulos, A.S., Pitsianis, N., Floros, D., Sismanis, N., Sun. X. Rapid Near-Neighbor Interactions in High Dimensions via Hierarchical Clustering [pdf]
AUTH Planned Thesis
  1. Bisbas, G. Implementing and improving point cloud simplification methods, MSc Thesis, 2019 [pdf]