Related Experiment Video
Updated: Mar 6, 2026
11:09
RBDT: A Computerized Task System based in Transposition for the Continuous Analysis of Relational Behavior Dynamics in Humans
Published on: July 17, 2021
3.4K
A database for reproducible manipulation research: CapriDB - Capture, Print, Innovate
Florian T Pokorny1, Yasemin Bekiroglu2, Karl Pauwels1
1KTH Royal Institute of Technology, Sweden.
Data in Brief
|March 15, 2017
Summary
This study introduces a new method for creating 3D object models using affordable cameras and 3D printing, enabling accessible robotic manipulation research. The CapriDB database provides 3D printable models and tracking features for complex object analysis.
Area of Science:
- Robotics
- Computer Vision
- 3D Modeling
Background:
- Creating 3D object databases for robotics typically requires expensive scanning equipment.
- Existing methods lack accessibility for researchers without specialized setups.
Purpose of the Study:
- To develop an inexpensive and accessible method for generating 3D object models for robotic manipulation research.
- To create a database of 3D printable models with tracking features.
Main Methods:
- Utilizing monocular or RGB-D camera images for 3D object model generation.
- Employing 3D printing to create physical replicas of the models.
- Integrating state-of-the-art object tracking algorithms.
Main Results:
- Generation of detailed, textured 3D mesh models from camera images.
- Creation of 3D printed replicas that closely approximate original objects.
- Development of CapriDB, an extensible database with 40 initial models and tracking features.
Conclusions:
- The proposed approach offers an affordable alternative to traditional 3D scanning for robotics.
- 3D printed objects allow precise control over physical properties for reproducible research.
- CapriDB facilitates adoption of this method for complex object manipulation studies.
