Publications

@inproceedings{peter_co-creating_2025,

title = {Co-creating a VR Narrative Experience of Constructing a Food Storage Following OvaHimba Traditional Practices},

author = {K. Peter and I. Makosa and S. Auala and L. Ndjao and D. Maasz and U. Mbinge and H. Winschiers-Theophilus},

url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-105007984089&doi=10.1145%2f3706370.3731652&partnerID=40&md5=36f95823413852d636b39bd561c97917},

doi = {10.1145/3706370.3731652},

isbn = {979-840071391-0 (ISBN)},

year  = {2025},

date = {2025-01-01},

booktitle = {IMX - Proc. ACM Int. Conf. Interact. Media Experiences},

pages = {418–423},

publisher = {Association for Computing Machinery, Inc},

abstract = {As part of an attempt to co-create a comprehensive virtual environment in which one can explore and learn traditional practices of the OvaHimba people, we have co-designed and implemented a VR experience to construct a traditional food storage. In collaboration with the OvaHimba community residing in Otjisa, we have explored culturally valid representations of the process. We have further investigated different techniques such as photogrammetry, generative AI and manual methods to develop 3D models. Our findings highlight the importance of context, process, and community-defined relevance in co-design, the fluidity of cultural realities and virtual representations, as well as technical challenges. © 2025 Copyright held by the owner/author(s).},

keywords = {3D Modelling, 3D models, 3d-modeling, Co-designs, Community-based, Community-Based Co-Design, Computer aided design, Cultural heritage, Cultural heritages, Food storage, Human computer interaction, Human engineering, Indigenous Knowledge, Information Systems, Interactive computer graphics, Interactive computer systems, IVR, Namibia, OvaHimba, Ovahimbum, Photogrammetry, Sustainable development, Virtual environments, Virtual Reality},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{wang_elevating_2023,

title = {Elevating Perception: Unified Recognition Framework and Vision-Language Pre-Training Using Three-Dimensional Image Reconstruction},

author = {Z. Wang and A. Joshi and G. Zhang and W. Ren and F. Jia and X. Sun},

url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85192837757&doi=10.1109%2fAIHCIR61661.2023.00105&partnerID=40&md5=0fe17cc622a9aa90e88b8c3e6a3bed3b},

doi = {10.1109/AIHCIR61661.2023.00105},

isbn = {979-835036036-3 (ISBN)},

year  = {2023},

date = {2023-01-01},

booktitle = {Proc. - Int. Conf. Artif. Intell., Human-Comput. Interact. Robot., AIHCIR},

pages = {592–596},

publisher = {Institute of Electrical and Electronics Engineers Inc.},

abstract = {This research project explores a paradigm shift in perceptual enhancement by integrating a Unified Recognition Framework and Vision-Language Pre-Training in three-dimensional image reconstruction. Through the synergy of advanced algorithms from computer vision & language processing, the project tries to enhance the precision and depth of perception in reconstructed images. This innovative approach holds the potential to revolutionize fields such as medical imaging, virtual reality, and computer-aided design, providing a comprehensive perspective on the intersection of multimodal data processing and perceptual advancement. The anticipated research outcomes are expected to significantly contribute to the evolution of technologies that rely on accurate and contextually rich three-dimensional reconstructions. Moreover, the research aims to reduce the constant need for new datasets by improving pattern recognition through 3D image patterning on backpropagation. This continuous improvement of vectors is envisioned to enhance the efficiency and accuracy of pattern recognition, contributing to the optimization of perceptual systems over time. © 2023 IEEE.},

keywords = {3D Model LLM, 3D modeling, 3D models, 3D Tech, 3d-modeling, Augmented Reality, Character recognition, Component, Computer aided design, Computer vision, Continuous time systems, Data handling, Generative AI, Image enhancement, Image Reconstruction, Image to Text Generation, Medical Imaging, Pattern recognition, Pre-training, Reconstructive Training, Text generations, Three dimensional computer graphics, Virtual Reality},

pubstate = {published},

tppubtype = {inproceedings}

}

@inproceedings{vaidhyanathan_spacify_2023,

title = {Spacify A Generative Framework for Spatial Comprehension, Articulation and Visualization using Large Language Models (LLMs) and eXtended Reality (XR)},

author = {V. Vaidhyanathan and T. R. Radhakrishnan and J. L. G. López},

editor = {Crawford A. and Diniz N.M. and Beckett R. and Vanucchi J. and Swackhamer M.},

url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85192831586&partnerID=40&md5=996906de0f5ef1e6c88b10bb65caabc0},

isbn = {979-898608059-8 (ISBN)},

year  = {2023},

date = {2023-01-01},

booktitle = {Habits Anthropocene: Scarcity Abundance Post-Mater. Econ. - Proc. Annu. Conf. Assoc. Comput. Aided Des. Archit., ACADIA},

volume = {2},

pages = {430–443},

publisher = {Association for Computer Aided Design in Architecture},

abstract = {Spatial design, the thoughtful planning and creation of built environments, typically requires advanced technical knowledge and visuospatial skills, making it largely exclusive to professionals like architects, interior designers, and urban designers. This exclusivity limits non-experts' access to spatial design, despite their ability to describe requirements and suggestions in natural language. Recent advancements in generative artificial intelligence (AI), particularly large language models (LLMs), and extended reality, (XR) offer the potential to address this limitation. This paper introduces Spacify (Figure 1), a framework that utilizes the generalizing capabilities of LLMs, 3D data-processing, and XR interfaces to create an immersive medium for language-driven spatial understanding, design, and visualization for non-experts. This paper describes the five components of Spacify: External Data, User Input, Spatial Interface, Large Language Model, and Current Spatial Design; which enable the use of generative AI models in a) question/ answering about 3D spaces with reasoning, b) (re)generating 3D spatial designs with natural language prompts, and c) visualizing designed 3D spaces with natural language descriptions. An implementation of Spacify is demonstrated via an XR smartphone application, allowing for an end-to-end, language-driven interior design process. User survey results from non-experts redesigning their spaces in 3D using this application suggest that Spacify can make spatial design accessible using natural language prompts, thereby pioneering a new realm of spatial design that is naturally language-driven. © ACADIA 2023. All rights reserved.},

keywords = {3D data processing, 3D spaces, Architectural design, Built environment, C (programming language), Computational Linguistics, Computer aided design, Computer architecture, Data handling, Data users, Data visualization, Immersive media, Interior designers, Language Model, Natural languages, Spatial design, Three dimensional computer graphics, Urban designers, User interfaces, Visualization},

pubstate = {published},

tppubtype = {inproceedings}

}