A Toolbox for Generating Multidimensional 3-D Objects with Fine-Controlled Feature Space: Quaddle 2.0

Area of Science:

• Computer Vision
• Human-Computer Interaction
• Cognitive Science

Background:

• Generating multidimensional 3D objects for research and gaming is challenging.
• Parametric control over visual features is crucial for applications like vision research and video games.

Purpose of the Study:

• To present a novel toolbox for efficient and controllable generation of multidimensional 3D rendered objects.
• To enable fine-tuning of object features and control over object similarity for research applications.

Main Methods:

• A Python-based toolbox was developed to control ten visual feature dimensions of 3D objects.
• The toolbox interfaces with Blender software for object creation and allows feature morphing.
• Objects are defined as multi-dimensional feature vectors, categorized into primary, secondary, and accessory dimensions.

Main Results:

• The toolbox enables automated generation of multiple 3D objects with controlled feature similarity.
• Objects can be generated in both 3D object and 2D image formats.
• Demonstrated application in a sequence learning task within an augmented reality environment using the Unity engine.

Conclusions:

• The developed toolbox facilitates efficient generation of multidimensional 3D objects with precise feature control.
• It supports the creation of objects with adjustable similarity, beneficial for visual cognitive research.
• The toolkit is valuable for developing immersive visual environments and advancing augmented reality applications.