This research aims to experiment and verify an innovative workflow for the construction of architectural-themed digital elements, such that thanks to the automatisms of image generation and virtualization it is particularly rapid compared to the canonical 3D modeling processes. The process starts from a series of textual keywords to arrive at interactive augmented reality simulations. The case study examined was that of the artistic and architectural styles respectively associated with Kazimir Malevič and Zaha Hadid. The latter is intimately linked in theoretical intentions to the Russian master, but has undertaken such a different artistic path that the hypotheses of visual mixes between the two styles generated with AI are particularly interesting. The analysis of these styles first allowed us to extract a series of keywords to insert into the AI text-to-image prompt, to create images of external and internal architectural environments, subsequently reconstructed in various compared attempts. The procedure that proved to be the most effective was that of semi-automatic perspective reconstruction, for relocating the elements of the image in digital space, thus creating a 3D model useful for equally quickly creating an augmented reality virtual environment, ready for use in a few minutes. The results obtained from the laboratory experimentation demonstrate the excellent effectiveness and speed of the workflow, proving the advantages of an integrative use of AI, which despite the imperfections offer very efficient solutions, as well as excellent food for thought on artistic paths that have not occurred in history.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

From Text to Augmented Reality: Experimentation of a Smart Workflow on the Works of Malevič and Hadid

  • Alessandro Martinelli,
  • Giulia Lazzaretto,
  • Agnese Toni,
  • Matteo Cavaglià

摘要

This research aims to experiment and verify an innovative workflow for the construction of architectural-themed digital elements, such that thanks to the automatisms of image generation and virtualization it is particularly rapid compared to the canonical 3D modeling processes. The process starts from a series of textual keywords to arrive at interactive augmented reality simulations. The case study examined was that of the artistic and architectural styles respectively associated with Kazimir Malevič and Zaha Hadid. The latter is intimately linked in theoretical intentions to the Russian master, but has undertaken such a different artistic path that the hypotheses of visual mixes between the two styles generated with AI are particularly interesting. The analysis of these styles first allowed us to extract a series of keywords to insert into the AI text-to-image prompt, to create images of external and internal architectural environments, subsequently reconstructed in various compared attempts. The procedure that proved to be the most effective was that of semi-automatic perspective reconstruction, for relocating the elements of the image in digital space, thus creating a 3D model useful for equally quickly creating an augmented reality virtual environment, ready for use in a few minutes. The results obtained from the laboratory experimentation demonstrate the excellent effectiveness and speed of the workflow, proving the advantages of an integrative use of AI, which despite the imperfections offer very efficient solutions, as well as excellent food for thought on artistic paths that have not occurred in history.