[ { "id": "rasoulzadeh-2025-archcomplete", "type_id": "journalpaper_notalk", "tu_id": null, "repositum_id": "20.500.12708/227557", "title": "ArchComplete: Autoregressive 3D architectural design generation with hierarchical diffusion-based upsampling", "date": "2025-12", "abstract": "Recent advances in 3D generative models have shown promising results but often fall short in capturing the complexity of architectural geometries and topologies. To tackle this, we present ArchComplete, a two-stage voxel-based 3D generative pipeline consisting of a vector-quantized model, whose composition is modeled with an autoregressive transformer for generating coarse shapes, followed by a set of multiscale diffusion models for augmenting with fine geometric details. Key to our pipeline is (i) learning a contextually rich codebook of local patch embeddings, optimized alongside a 2.5D perceptual loss that captures global spatial correspondence of projections onto three axis-aligned orthogonal planes, and (ii) redefining upsampling as a set of multiscale conditional diffusion models learning over a hierarchy of coarse-to-fine local volumetric patches, with a guided denoising process using 3D Gaussian windows that smooths noise estimates across overlapping patches during inference. Trained on our introduced dataset of 3D house models, ArchComplete autoregressively generates models at the resolution of (Formula presented) and progressively refines them up to (Formula presented), with voxel sizes as small as (Formula presented). ArchComplete solves a variety of tasks, including genetic interpolation and variation, unconditional synthesis, shape and plan-drawing completion, as well as geometric detailization, while achieving state-of-the-art performance.", "authors_et_al": false, "substitute": null, "main_image": null, "sync_repositum_override": null, "repositum_presentation_id": null, "authors": [ 5233, 5594, 1799, 5203, 5204, 193 ], "articleno": "104477", "doi": "10.1016/j.cag.2025.104477", "issn": "1873-7684", "journal": "COMPUTERS & GRAPHICS-UK", "pages": "12", "publisher": "PERGAMON-ELSEVIER SCIENCE LTD", "volume": "133", "research_areas": [], "keywords": [ "Architectural Geometries", "Geometric Deep Learning", "Multiresolution Modeling", "Shape Analysis and Synthesis", "Shape Modeling Applications" ], "weblinks": [], "files": [], "projects_workgroups": [ "d4314" ], "url": "https://www.cg.tuwien.ac.at/research/publications/2025/rasoulzadeh-2025-archcomplete/", "__class": "Publication" }, { "id": "rasoulzadeh-2024-strokes2surface", "type_id": "journalpaper_notalk", "tu_id": null, "repositum_id": "20.500.12708/208007", "title": "Strokes2Surface: Recovering Curve Networks From 4D Architectural Design Sketches", "date": "2024-05", "abstract": "We present Strokes2Surface, an offline geometry reconstruction pipeline that recovers well-connected curve networks from imprecise 4D sketches to bridge concept design and digital modeling stages in architectural design. The input to our pipeline consists of 3D strokes' polyline vertices and their timestamps as the 4th dimension, along with additional metadata recorded throughout sketching. Inspired by architectural sketching practices, our pipeline combines a classifier and two clustering models to achieve its goal. First, with a set of extracted hand-engineered features from the sketch, the classifier recognizes the type of individual strokes between those depicting boundaries (Shape strokes) and those depicting enclosed areas (Scribble strokes). Next, the two clustering models parse strokes of each type into distinct groups, each representing an individual edge or face of the intended architectural object. Curve networks are then formed through topology recovery of consolidated Shape clusters and surfaced using Scribble clusters guiding the cycle discovery. Our evaluation is threefold: We confirm the usability of the Strokes2Surface pipeline in architectural design use cases via a user study, we validate our choice of features via statistical analysis and ablation studies on our collected dataset, and we compare our outputs against a range of reconstructions computed using alternative methods.", "authors_et_al": false, "substitute": null, "main_image": null, "sync_repositum_override": null, "repositum_presentation_id": null, "authors": [ 5233, 193, 5429, 1799 ], "articleno": "e15054", "doi": "10.1111/cgf.15054", "issn": "1467-8659", "journal": "Computer Graphics Forum", "number": "2", "pages": "16", "pages_from": "1", "pages_to": "16", "publisher": "WILEY", "volume": "43", "research_areas": [], "keywords": [ "CCS Concepts", "Computer graphics", "Computing methodologies → Artificial intelligence", "Machine learning" ], "weblinks": [], "files": [], "projects_workgroups": [ "d4314" ], "url": "https://www.cg.tuwien.ac.at/research/publications/2024/rasoulzadeh-2024-strokes2surface/", "__class": "Publication" }, { "id": "reisinger-2023-iad", "type_id": "inproceedings", "tu_id": null, "repositum_id": "20.500.12708/188467", "title": "Integrating AEC Domain-Specific Multidisciplinary Knowledge for Informed and Interactive Feedback in Early Design Stages", "date": "2023-10", "abstract": "In the context of digitalization in the industry, a variety of technologies has been developed for system integration and enhanced team collaboration in the Architecture, Engineering and Construction (AEC) industry. Multidisciplinary design requirements are characterized by a high degree of complexity. Early design methods often rely on implicit or experiential design knowledge, whereas contemporary digital design tools mostly reflect domain-specific silo thinking with time-consuming iterative design processes. Yet, the early design stages hold the greatest potential for design optimization. This paper presents a framework of a multidisciplinary computational integration platform for early design stages that enables integration of AEC domain-specific methods from architecture, engineering, mathematics and computer science. The platform couples a semantic integrative mixed reality sketching application to a shape inference machine-learning based algorithm to link methods for different computation, simulation and digital fabrication tasks. A proof of concept of the proposed framework is presented for the use case of a freeform geometry wall. Future research will explore the potential of the framework to be extended to larger building projects with the aim to connect the method into BIM-processes.", "authors_et_al": false, "substitute": null, "main_image": null, "sync_repositum_override": "date", "repositum_presentation_id": null, "authors": [ 1874, 5233, 1487, 240, 1712, 5202, 1799, 193 ], "booktitle": "Advances in Information Technology in Civil and Building Engineering: Proceedings of ICCCBE 2022 - Volume 2", "date_from": "2022-10-26", "date_to": "2022-10-28", "doi": "10.1007/978-3-031-32515-1_12", "event": "19th International Conference on Computing in Civil and Building Engineering (ICCCBE 2022)", "isbn": "978-3-031-32515-1", "lecturer": [ 1874 ], "location": "Cape Town", "pages": "18", "pages_from": "153", "pages_to": "170", "publisher": "Springer", "volume": "358", "research_areas": [ "Modeling" ], "keywords": [ "Integrated Design", "Early Design Stage", "Mixed Reality Sketching", "Shape Inference", "Computational Design", "Integration Platform", "Digital Fabrication" ], "weblinks": [], "files": [], "projects_workgroups": [ "d4314" ], "url": "https://www.cg.tuwien.ac.at/research/publications/2023/reisinger-2023-iad/", "__class": "Publication" }, { "id": "rasoulzadeh-2023-ani", "type_id": "journalpaper_notalk", "tu_id": null, "repositum_id": "20.500.12708/189820", "title": "A Novel Integrative Design Framework Combining 4D Sketching, Geometry Reconstruction, Micromechanics Material Modelling, and Structural Analysis", "date": "2023-08", "abstract": "State-of-the-art workflows within Architecture, Engineering, and Construction (AEC) are still caught in sequential planning processes. Digital design tools in this domain often lack proper communication between different stages of design and relevant domain knowledge. Furthermore, decisions made in the early stages of design, where sketching is used to initiate, develop, and communicate ideas, heavily impact later stages, resulting in the need for rapid feedback to the architectural designer so they can proceed with adequate knowledge about design implications. Accordingly, this paper presents research on a novel integrative design framework based on a recently developed 4D sketching interface, targeted for architectural design as a form-finding tool coupled with three modules: (1) a Geometric Modelling module, which utilises Points2Surf as a machine learning model for automatic surface mesh reconstruction from the point clouds produced by sketches, (2) a Material Modelling module, which predicts the mechanical properties of biocomposites based on multiscale micromechanics homogenisation techniques, and (3) a Structural Analysis module, which assesses the mechanical performance of the meshed structure on the basis of the predicted material properties using finite element simulations. The proposed framework is a step towards using material-informed design already in the early stages of design.", "authors_et_al": false, "substitute": null, "main_image": null, "sync_repositum_override": null, "repositum_presentation_id": null, "authors": [ 5233, 5303, 5304, 1874, 1799, 5209, 193 ], "articleno": "102074", "doi": "10.1016/j.aei.2023.102074", "issn": "1873-5320", "journal": "Advanced Engineering Informatics", "publisher": "ELSEVIER SCI LTD", "volume": "57", "research_areas": [], "keywords": [ "3D reconstruction", "Biocomposite", "Early-design stage", "Finite element analysis", "Machine learning", "Material-informed", "Micromechanics", "Multiscale modelling", "Sketch-based interface", "Sketch-based modelling", "Structural analysis" ], "weblinks": [], "files": [], "projects_workgroups": [], "url": "https://www.cg.tuwien.ac.at/research/publications/2023/rasoulzadeh-2023-ani/", "__class": "Publication" }, { "id": "rasoulzadeh-2022-strokes2surface", "type_id": "talk", "tu_id": null, "repositum_id": "20.500.12708/153311", "title": "Strokes2Surface: Recovering Curve Networks from 4D Architectural Design Sketches for Shape Inference", "date": "2022-10-12", "abstract": null, "authors_et_al": false, "substitute": null, "main_image": null, "sync_repositum_override": null, "repositum_presentation_id": null, "authors": [ 5233, 193, 1799 ], "date_from": "2022-10-10", "date_to": "2022-10-12", "event": "Advance AEC Autumn School", "lecturer": [ 5233 ], "research_areas": [], "keywords": [ "Architectural Geometry", "Concept Design", "Digital Modeling", "Machine Learning", "Curve Networks", "Shape Inference" ], "weblinks": [], "files": [], "projects_workgroups": [ "d4314" ], "url": "https://www.cg.tuwien.ac.at/research/publications/2022/rasoulzadeh-2022-strokes2surface/", "__class": "Publication" } ]