Cumincad : CUMINCAD Papers : Search Results

CumInCAD is a Cumulative Index about publications in Computer Aided Architectural Design
supported by the sibling associations ACADIA, CAADRIA, eCAADe, SIGraDi, ASCAAD and CAAD futures

PDF papers
References

Hits 1 to 20 of 2736

Reformat results as: short short into frame detailed detailed into frame

_id	ascaad2021_093
id	ascaad2021_093
authors	Alani, Mostafa; Bilal Al-Kaseem
year	2021
title	Fill in the Blanks: Deep Convolutional Generative Adversarial Networks to Investigate the Virtual Design Space of Historical Islamic Patterns
source	Abdelmohsen, S, El-Khouly, T, Mallasi, Z and Bennadji, A (eds.), Architecture in the Age of Disruptive Technologies: Transformations and Challenges [9th ASCAAD Conference Proceedings ISBN 978-1-907349-20-1] Cairo (Egypt) [Virtual Conference] 2-4 March 2021, pp. 614-621
summary	This paper presents a method to explore the virtual design space of historical Islamic Geometric Patterns (IGP). The introduced approach utilizes Deep Convolutional Generative Adversarial Network (DCGAN) to learn from historically existing hexagonal-based IGP to synthesis novel, authentically looking Geometric Patterns.
series	ASCAAD
email
last changed	2021/08/09 13:13

_id	ascaad2021_022
id	ascaad2021_022
authors	Baºarir, Lale; Kutluhan Erol
year	2021
title	Briefing AI: From Architectural Design Brief Texts to Architectural Design Sketches
source	Abdelmohsen, S, El-Khouly, T, Mallasi, Z and Bennadji, A (eds.), Architecture in the Age of Disruptive Technologies: Transformations and Challenges [9th ASCAAD Conference Proceedings ISBN 978-1-907349-20-1] Cairo (Egypt) [Virtual Conference] 2-4 March 2021, pp. 23-31
summary	The main focus of this research is to uncover the underlying intuitive knowledge of architecture with the help of machine learning models. To achieve this, a generic architectural design process is considered and divided into iterative portions based on their output for each phase. This study looks into the initial portion of the architectural design process called “Briefing”. The authors search for the intuition that exists within the design process and how it can be learned by artificial intelligence (AI) that is currently gained through master-apprentice relationship and experience that builds up this knowledge. In this study, a way to enable users to attain an architectural design sketch while defining an architectural design problem with text is explored. This on-going research decomposes the components of the briefing and preliminary design sketching processes. Therefore the domain knowledge at each phase is considered for translating to constraints via natural language processing (NLP) and machine learning (ML) models such as Generative Adversarial Networks (GANs).
series	ASCAAD
type	normal paper
email
last changed	2021/08/09 13:11

_id	caadria2021_225
id	caadria2021_225
authors	Cao, Shuqi and Ji, Guohua
year	2021
title	Automatically generating layouts of large-scale office park using position-based dynamics
source	A. Globa, J. van Ameijde, A. Fingrut, N. Kim, T.T.S. Lo (eds.), PROJECTIONS - Proceedings of the 26th CAADRIA Conference - Volume 1, The Chinese University of Hong Kong and Online, Hong Kong, 29 March - 1 April 2021, pp. 21-30
doi	https://doi.org/10.52842/conf.caadria.2021.1.021
summary	In this paper we propose an automatic layout algorithm using PBD (Position-Based Dynamic) for large-scale office park planning. Typically, the organization of buildings into a layout is a labor-intensive problem, and takes up most of designers working time. Unlike Evolutionary Algorithms who has high computational cost, and GAN (Generative Adversarial Networks) whose constraints are not explicit, PBD can handle complex geometric constraints fast enough to be used in interactive environments. The high efficiency will not only accelerate the design iteration from draft to drawings, but also provide precious feasible sample for performance optimization. Furthermore, PBD is intuitive and flexible to be implemented which makes it a potential technique to be used in real design workflow.
keywords	Generative Design; Automated Layout Generation; Position-Based Dynamics; Real-time Design Tool; Exploratory Design
series	CAADRIA
email
last changed	2022/06/07 07:54

_id	sigradi2021_114
id	sigradi2021_114
authors	Cesar Rodrigues, Ricardo, Kenzo Imagawa, Marcelo, Rubio Koga, Renan and Bertola Duarte, Rovenir
year	2021
title	Big Data vs Smart Data on the Generation of Floor Plans with Deep Learning
source	Gomez, P and Braida, F (eds.), Designing Possibilities - Proceedings of the XXV International Conference of the Ibero-American Society of Digital Graphics (SIGraDi 2021), Online, 8 - 12 November 2021, pp. 217–228
summary	Due to the progressive growth of data dimensionality, addressing how much data and time is required to train deep learning models has become an important research topic. Thus, in this paper, we present a benchmark for generating floor plans with Conditional Generative Adversarial Networks in which we compare 10 trained models on a dataset of 80.000 samples, the models use different data dimensions and hyper-parameters on the training phase, beyond this objective, we also tested the capability of Convolutional Neural Networks (CNN) to reduce the dataset noise. The models' assessment was made on more than 6 million with the Frétche Inception Distance (FID). The results show that such models can rapidly achieve similar or even better FID results if trained with 800 images of 512x512 pixels, in comparison to high dimensional datasets of 256x256 pixels, however, using CNNs to enhance data consistency reproduced optimal results using around 27.000 images.
keywords	Floor plans, Generative design, Generative adversarial networks, Smart Data, Dataset reduction.
series	SIGraDi
email
last changed	2022/05/23 12:10

_id	caadria2021_389
id	caadria2021_389
authors	del Campo, Matias
year	2021
title	Architecture,Language and AI - Language,Attentional Generative Adversarial Networks (AttnGAN) and Architecture Design
source	A. Globa, J. van Ameijde, A. Fingrut, N. Kim, T.T.S. Lo (eds.), PROJECTIONS - Proceedings of the 26th CAADRIA Conference - Volume 1, The Chinese University of Hong Kong and Online, Hong Kong, 29 March - 1 April 2021, pp. 211-220
doi	https://doi.org/10.52842/conf.caadria.2021.1.211
summary	The motivation to explore Attentional Generative Adversarial Networks (AttnGAN) as a design technique in architecture can be found in the desire to interrogate an alternative design methodology that does not rely on images as starting point for architecture design, but language. Traditionally architecture design relies on visual language to initiate a design process, wither this be a napkin sketch or a quick doodle in a 3D modeling environment. AttnGAN explores the information space present in programmatic needs, expressed in written form, and transforms them into a visual output. The key results of this research are shown in this paper with a proof-of-concept project: the competition entry for the 24 Highschool in Shenzhen, China. This award-winning project demonstrated the ability of GraphCNN to serve as a successful design methodology for a complex architecture program. In the area of Neural Architecture, this technique allows to interrogate shape through language. An alternative design method that creates its own unique sensibility.
keywords	Artificial Intelligence; Machine Learning; Artificial Neural Networks; Semiotics; Design Methodology
series	CAADRIA
email
last changed	2022/06/07 07:55

_id	ecaade2021_109
id	ecaade2021_109
authors	Doumpioti, Christina and Huang, Jeffrey
year	2021
title	Intensive Differences in Spatial Design - Reversing form-finding
source	Stojakovic, V and Tepavcevic, B (eds.), Towards a new, configurable architecture - Proceedings of the 39th eCAADe Conference - Volume 1, University of Novi Sad, Novi Sad, Serbia, 8-10 September 2021, pp. 9-16
doi	https://doi.org/10.52842/conf.ecaade.2021.1.009
summary	Drawing from the philosophy of science, 'intensive' qualities define differences in degree instead of 'extensive' ones that define additive quantities. More relevant to architecture, intensive differences can define transient boundaries such as warmness and coolness, dryness and moisture, light and shadow, or visual accessibility, to name a few.The question that serves as a starting point of this study is whether the attributes mentioned above can become form-giving agents during the design process and, therefore, whether they become fundamental parameters for the conceptualization and configuration of extensive spatial qualities. This question is explored using Generative Adversarial Networks and image-to-image translation. The dataset consists of two types of images; one consists of spatial configurations representing extensive attributes. The second set depicts intensive characteristics of visual accessibility. The study proposes a conceptual model and workflow that reverses form-finding and enables the design of environments through the specification of desired intensive attributes. Furthermore, it discusses the advantage of working with this method in search of architectural environments with embedded spatial experiences.
keywords	Intensive Differences; Form-Finding; Isovist Simulation; conditional Generative Adversarial Networks (cGAN)
series	eCAADe
email
last changed	2022/06/07 07:55

_id	cdrf2021_3
id	cdrf2021_3
authors	Jean Jaminet, Gabriel Esquivel, and Shane Bugni
year	2021
title	Serlio and Artificial Intelligence: Problematizing the Image-to-Object Workflow
source	Proceedings of the 2021 DigitalFUTURES The 3rd International Conference on Computational Design and Robotic Fabrication (CDRF 2021)
doi	https://doi.org/https://doi.org/10.1007/978-981-16-5983-6_1
summary	Virtual design production demands that information be increasingly encoded and decoded with image compression technologies. Since the Renaissance, the discourses of language and drawing and their actuation by the classical disciplinary treatise have been fundamental to the production of knowledge within the building arts. These early forms of data compression provoke reflection on theory and technology as critical counterparts to perception and imagination unique to the discipline of architecture. This research examines the illustrated expositions of Sebastiano Serlio through the lens of artificial intelligence (AI). The mimetic powers of technological data storage and retrieval and Serlio’s coded operations of orthographic projection drawing disclose other aesthetic and formal logics for architecture and its image that exist outside human perception. Examination of aesthetic communication theory provides a conceptual dimension of how architecture and artificial intelligent systems integrate both analog and digital modes of information processing. Tools and methods are reconsidered to propose alternative AI workflows that complicate normative and predictable linear design processes. The operative model presented demonstrates how augmenting and interpreting layered generative adversarial networks drive an integrated parametric process of three-dimensionalization. Concluding remarks contemplate the role of human design agency within these emerging modes of creative digital production.
series	cdrf
email
last changed	2022/09/29 07:53

_id	ecaade2021_158
id	ecaade2021_158
authors	Joyce, Sam Conrad and Nazim, Ibrahim
year	2021
title	Limits to Applied ML in Planning and Architecture - Understanding and defining extents and capabilities
source	Stojakovic, V and Tepavcevic, B (eds.), Towards a new, configurable architecture - Proceedings of the 39th eCAADe Conference - Volume 1, University of Novi Sad, Novi Sad, Serbia, 8-10 September 2021, pp. 243-252
doi	https://doi.org/10.52842/conf.ecaade.2021.1.243
summary	There has been an exponential increase in Machine Learning (ML) research in design. Specifically, with Deep Learning becoming more accessible, frameworks like Generative Adversarial Networks (GANs), which are able to synthesise novel images are being used in the classification and generation of designs in architecture. While much of these explorations successfully demonstrate the 'magic' and potential of these techniques, their limits remain unclear, with only a few, but crucial, discussions on underlying fundamental limits and sensitivities of ML. This is a gap in our understanding of these tools especially within the complex context of planning and architecture. This paper seeks to discuss what limits ML in design as it exists today, by examining the state-of-the-art and mechanics of ML models relevant to design tasks. Aiming to help researchers to focus on productive uses of ML and avoid areas of over-promise.
keywords	Machine Learning; Artificial Intelligence; Creativity
series	eCAADe
email
last changed	2022/06/07 07:52

_id	acadia21_112
id	acadia21_112
authors	Kahraman, Ridvan; Zechmeister, Christoph; Dong, Zhetao; Oguz, Ozgur S.; Drachenberg, Kurt; Menges, Achim; Rinderspacher, Katja
year	2021
title	Augmenting Design
source	ACADIA 2021: Realignments: Toward Critical Computation [Proceedings of the 41st Annual Conference of the Association of Computer Aided Design in Architecture (ACADIA) ISBN 979-8-986-08056-7]. Online and Global. 3-6 November 2021. edited by B. Bogosian, K. Dörfler, B. Farahi, J. Garcia del Castillo y López, J. Grant, V. Noel, S. Parascho, and J. Scott. 112-121.
doi	https://doi.org/10.52842/conf.acadia.2021.112
summary	In recent years, generative machine learning methods such as variational autoencoders (VAEs) and generative adversarial networks (GANs) have opened up new avenues of exploration for architects and designers. The presented work explores how these methods can be expanded by incorporating multiple abstract criteria directly into the formulation of the algorithm that negotiates these complex criteria and proposes a fitting design. It draws inspiration from the works of several design theorists who have developed such goal-oriented approaches to design, and sets up multiple-objective VAE and GAN frameworks with this idea in mind. The research demonstrates that by incorporating multiple constraints using auxiliary discriminator networks, the developed algorithms are able to generate innovative solutions to two example problems: the design of 2D digits, and the design of 3D voxel chairs. By speculating and examining the role of the designer in data based generative computational design workflows, the research aims to provide an approach for solving design tasks in the age of big data.
series	ACADIA
type	paper
email
last changed	2023/10/22 12:06

_id	sigradi2021_200
id	sigradi2021_200
authors	Karabagli, Kaan, Koc, Mustafa, Basu, Prithwish and As, Imdat
year	2021
title	A Machine Learning Approach to Translate Graph Representations into Conceptual Massing Models
source	Gomez, P and Braida, F (eds.), Designing Possibilities - Proceedings of the XXV International Conference of the Ibero-American Society of Digital Graphics (SIGraDi 2021), Online, 8 - 12 November 2021, pp. 191–202
summary	Machine learning (ML) has popular applications in domains involving image, video, text and voice. However, in architecture, image-based ML systems face challenges capturing the complexity of three-dimensional space. In this paper, we leverage a graph-based ML system that can capture the inherent topology of architectural conceptual designs and identify high-performing latent patterns within such designs. In particular, our goal is to translate architectural graph data into three-dimensional massing models. We are building on our prior ML work, where we, a. discovered latent topological features, b. composed building blocks into new designs, c. evaluated their feasibility, and d. explored Generative Adversarial (Neural) Networks (GAN)-generated design variations. We trained the ML system with architectural design data that we gathered from an online architectural design competition platform, translated them into machine-readable graph representations, and identified their essential subgraphs to develop novel compositions. In this paper, we explore how these novel designs (outputted in graph form), can be translated into three-dimensional architectural form. We present an ML approach to turn graph representations into functional volumetric massing models. The ultimate goal of the study is to develop an end-to-end pipeline to generate architectural design - from a graph representation to a fully developed conceptual proxy of a designed product. The research question is promising in automating conceptual design, and we believe the outcome can be relevant to other design disciplines as well.
keywords	Architectural design, machine learning, conceptual design, deep learning, artificial intelligence
series	SIGraDi
email
last changed	2022/05/23 12:10

_id	ecaade2021_037
id	ecaade2021_037
authors	Kikuchi, Takuya, Fukuda, Tomohiro and Yabuki, Nobuyoshi
year	2021
title	Automatic Diminished Reality-Based Virtual Demolition Method using Semantic Segmentation and Generative Adversarial Network for Landscape Assessment
source	Stojakovic, V and Tepavcevic, B (eds.), Towards a new, configurable architecture - Proceedings of the 39th eCAADe Conference - Volume 2, University of Novi Sad, Novi Sad, Serbia, 8-10 September 2021, pp. 529-538
doi	https://doi.org/10.52842/conf.ecaade.2021.2.529
summary	In redevelopment projects in mature cities, it is important to visualize the future landscape. Diminished reality (DR) based methods have been proposed to represent the future landscape after the structures are removed. However, two issues remain to be addressed in previous studies. (1) the user needs to prepare 3D models of the structure to be removed and the background structure to be rendered after removal as preprocessing, and (2) the user needs to specify the structure to be removed in advance. In this study, we propose a DR method that detects the objects to be removed using semantic segmentation and completes the removal area using generative adversarial networks. With this method, virtual removal can be performed without preparing 3D models in advance and without specifying the removal target in advance. A prototype system was used for verification, and it was confirmed that the method can represent the future landscape after removal and can run at an average speed of about 8.75 fps.
keywords	landscape visualization; virtual demolition; diminished reality (DR); deep learning; generative adversarial network (GAN); semantic segmentation
series	eCAADe
email
last changed	2022/06/07 07:52

_id	sigradi2021_5
id	sigradi2021_5
authors	Ng, Provides, Fernandez, Alberto, Doria, David, Odaibat, Baha and Karastathi, Nikoletta
year	2021
title	AI In+form: Intelligence and Aggregation for Solar Designs in the Built Environment
source	Gomez, P and Braida, F (eds.), Designing Possibilities - Proceedings of the XXV International Conference of the Ibero-American Society of Digital Graphics (SIGraDi 2021), Online, 8 - 12 November 2021, pp. 203–215
summary	Designers are increasingly challenged by a constant change of context and the interaction of layers of data from a huge variety of sources, from natural-artificial to human-machine. This research aims at mapping the interrelations of energy problems, bio- and artificial intelligence, and human-machine interaction to reflect and rethink the future of solar design. This paper first discusses its theoretical approach that stands at the convergence of light-harvesting systems, their aggregation and intelligence. Afterwhich, this paper explores their translation into iterative processes between designer and artificial intelligences, which is defined as rule/agent-based and machine learning systems; in particular, the relationship between Cellular Automata, Genetic Algorithm, and Generative Adversarial Networks (GANs) is discussed. Finally, it introduces a design project - @R.E.Ar_ - showing the proposed combinatorial pipeline and some preliminary results.
keywords	artificial intelligence, bio-inspired, solar design, Aggregation, human-machine interaction
series	SIGraDi
email
last changed	2022/05/23 12:10

_id	caadria2021_043
id	caadria2021_043
authors	Ng, Provides
year	2021
title	21E8: Coupling Generative Adversarial Neural Networks (GANS) with Blockchain Applications in Building Information Modelling (BIM) Systems
source	A. Globa, J. van Ameijde, A. Fingrut, N. Kim, T.T.S. Lo (eds.), PROJECTIONS - Proceedings of the 26th CAADRIA Conference - Volume 2, The Chinese University of Hong Kong and Online, Hong Kong, 29 March - 1 April 2021, pp. 111-120
doi	https://doi.org/10.52842/conf.caadria.2021.2.111
summary	The ability of GANs to synthesize large sets of data is ideal for coupling with BIM to formulate a multi-access system that enables users to search and browse through a spectrum of articulated options, all personalised to design specificity - an 'Architecture Machine'. Nonetheless, due to challenges in proprietary incompatibility, BIM systems currently lack a secured yet transparent way of freely integrating with crowdsourced efforts. This research proposes to employ blockchain as a means to couple GANs and BIM, with e8 networking topology to facilitate communication and distribution. It consists of a literature review and a design research that proposes a tech stack design and UML (unified modeling language) use cases, and presents preliminary design results obtained using GANs and e8.
keywords	21e8; GANs; Blockchain; BIM; Architecture Machine
series	CAADRIA
email
last changed	2022/06/07 07:58

_id	ecaade2021_290
id	ecaade2021_290
authors	Nicholas, Paul, Chen, Yu, Borpujari, Nihit, Bartov, Nitsan and Refsgaard, Andreas
year	2021
title	A Chained Machine Learning Approach to Motivate Retro-Cladding of Residential Buildings
source	Stojakovic, V and Tepavcevic, B (eds.), Towards a new, configurable architecture - Proceedings of the 39th eCAADe Conference - Volume 1, University of Novi Sad, Novi Sad, Serbia, 8-10 September 2021, pp. 55-64
doi	https://doi.org/10.52842/conf.ecaade.2021.1.055
summary	This paper investigates how a novel approach to visualisation could help address the challenge of motivating residential retrofitting. Emerging retrofitting research and practice emphasises retro-cladding - the upgrading of the exterior facade of a building - using a modular approach. We present a machine-learning based approach aimed to motivate residential retrofitting through the generation of images and cost/benefit information describing climatically specific additions of external insulation and green roof panels to the façade of a Danish type house. Our approach chains a series of different models together, and implements a method for the controlled navigation of the principle generative styleGAN model. The approach is at a prototypical stage that implements a full workflow but does not include numerical evaluation of model predictions. Our paper details our processes and considerations for the generation of new datasets, the specification and chaining of models, and the linking of climatic data to travel through the latent space of a styleGAN model to visualise and provide a simple cost benefit report for retro-cladding specific to the local climates of five different Danish cities.
keywords	Retrofitting; Machine Learning; Generative Adversarial Networks; Synthetic Datasets
series	eCAADe
type	normal paper
email
last changed	2022/06/07 07:58

_id	caadria2021_191
id	caadria2021_191
authors	Shou, Xinyue, Chen, Pinyang and Zheng, Hao
year	2021
title	Predicting the Heat Map of Street Vendors from Pedestrian Flow through Machine Learning
source	A. Globa, J. van Ameijde, A. Fingrut, N. Kim, T.T.S. Lo (eds.), PROJECTIONS - Proceedings of the 26th CAADRIA Conference - Volume 2, The Chinese University of Hong Kong and Online, Hong Kong, 29 March - 1 April 2021, pp. 569-578
doi	https://doi.org/10.52842/conf.caadria.2021.2.569
summary	Street vending is a recent policy advocated by city governments to support small and intermediate businesses in the post-pandemic period in China. Street vendors select their locations primarily based on their intuitions about the surrounding environment; they temporarily occupy popular locations that benefit their business. Taking the city of Chengdu as an example, this study aims to formulate the rules governing vendors location selection using machine learning and big data analysis techniques, thus identifying streets likely to become vital street markets. We propose a semantic segmentation method to construct heat maps that visualize and quantify the distribution of street vendors and pedestrians on public urban streets. The image-based generative adversarial network (GAN) is then trained to predict the vendors heat maps from the pedestrians heat map, finding the relationship between the locations of the vendors and the pedestrians. Our successful prediction of the vendors locations highlights machine learning techniques ability to quantify experience-based decision strategies. Moreover, suggesting potential marketing locations to vendors could help increase cities vitality.
keywords	Machine Learning; Big Data Analysis; Semantic Segmentation; Generative Adversarial Networks
series	CAADRIA
email
last changed	2022/06/07 07:56

_id	sigradi2021_134
id	sigradi2021_134
authors	Uzun, Can
year	2021
title	What can Colors and Shapes Tell about Generative Adversarial Networks?
source	Gomez, P and Braida, F (eds.), Designing Possibilities - Proceedings of the XXV International Conference of the Ibero-American Society of Digital Graphics (SIGraDi 2021), Online, 8 - 12 November 2021, pp. 161–171
summary	The study aims to understand the how’s and what’s of creating an architectural dataset for generative adversarial nets through the evaluation of the effects of colors and shapes in image datasets on generative adversarial nets. Throughout the paper, six generative adversarial network training sessions are conducted on DCGAN and context-encoder algorithms with three different datasets having different complexities for colors and shapes. Firstly the color and shape complexities are analyzed for datasets. For color complexity, heuristic analyze is applied and for shape complexity, gray level occurrence matrix entropy which gives the textural complexity is utilized. In the end, the complexities and the training results are evaluated. Results show that color complexity has an important role for generative adversarial networks to generate colors correctly. Regularity in shape complexity /gray level co-occurrence matrix entropy distribution facilitates the algorithm training and shape generating processes.
keywords	Context-Encoder, GAN, Colors, Shapes
series	SIGraDi
email
last changed	2022/05/23 12:10

_id	cdrf2021_69
id	cdrf2021_69
authors	Virginia Ellyn Melnyk
year	2021
title	Punch Card Patterns Designed with GAN
source	Proceedings of the 2021 DigitalFUTURES The 3rd International Conference on Computational Design and Robotic Fabrication (CDRF 2021)
doi	https://doi.org/https://doi.org/10.1007/978-981-16-5983-6_7
summary	Knitting punch cards codify different stitch patterns into binary patterns, telling the machine when to change color or to generate different stitch types. This research utilizes Neural Networks (NN) and image-based Generative Adversarial Networks (GAN), with an image database of knitting punch cards, to generate new punch card designs. The hypothesis is that artificial intelligence will learn the basic underlying structures of the punch cards and the pattern makeup that is inherent across patterns of different styles and cultures. Different neural networks were utilized throughout the research, such as Neural Style Transfer (NST), AdaIN Style Transfers, and StyleGAN2. The results from these explorations offer different insights into pattern design and various outcomes of the different neural networks. Ultimately physically testing these punch card designs, these patterns were knit on a domestic knitting machine, resulting in novel fabrication and design techniques that are both digital and craft-based.
series	cdrf
email
last changed	2022/09/29 07:53

_id	cdrf2021_26
id	cdrf2021_26
authors	Yuqian Li and Weiguo Xu
year	2021
title	Using CycleGAN to Achieve the Sketch Recognition Process of Sketch-Based Modeling
source	Proceedings of the 2021 DigitalFUTURES The 3rd International Conference on Computational Design and Robotic Fabrication (CDRF 2021)
doi	https://doi.org/https://doi.org/10.1007/978-981-16-5983-6_3
summary	Architects usually design ideation and conception by hand-sketching. Sketching is a direct expression of the architect’s creativity. But 2D sketches are often vague, intentional and even ambiguous. In the research of sketch-based modeling, it is the most difficult part to make the computer to recognize the sketches. Because of the development of artificial intelligence, especially deep learning technology, Convolutional Neural Networks (CNNs) have shown obvious advantages in the field of extracting features and matching, and Generative Adversarial Neural Networks (GANs) have made great breakthroughs in the field of architectural generation which make the image-to-image translation become more and more popular. As the building images are gradually developed from the original sketches, in this research, we try to develop a system from the sketches to the images of buildings using CycleGAN algorithm. The experiment demonstrates that this method could achieve the mapping process from the sketches to images, and the results show that the sketches’ features could be recognised in the process.By the learning and training process of the sketches’ reconstruction, the features of the images are also mapped to the sketches, which strengthen the architectural relationship in the sketch, so that the original sketch can gradually approach the building images, and then it is possible to achieve the sketch-based modeling technology.
series	cdrf
email
last changed	2022/09/29 07:53

_id	caadria2021_415
id	caadria2021_415
authors	Chuang, Cheng-Lin and Chien, Sheng-Fen
year	2021
title	Facilitating Architect-Client Communication in the Pre-design Phase
source	A. Globa, J. van Ameijde, A. Fingrut, N. Kim, T.T.S. Lo (eds.), PROJECTIONS - Proceedings of the 26th CAADRIA Conference - Volume 2, The Chinese University of Hong Kong and Online, Hong Kong, 29 March - 1 April 2021, pp. 71-80
doi	https://doi.org/10.52842/conf.caadria.2021.2.071
summary	The process of architects exploring the program with clients often take place through face-to-face oral discussions and visual aids, such as photos and sketches. Our research focuses on two communication mediums: language and sketch. We employ machine learning techniques to assist architects and clients to improve their communication and reduce misunderstandings. We have trained a Naive Bayesian Classifier machine, the language assistant (LA), to classify architectural vocabularies with associations to design requirements. In addition, we have trained a Generative Adversarial Network, the sketch assistant (SA), to generate photo quality images based on architects' sketches. The language assistant and sketch assistant combined can facilitate architect-client communication during the pre-design stage.
keywords	Architect-Client Communication; Pre-design; Architectural Programming; Machine Learning; Schematic Design
series	CAADRIA
email
last changed	2022/06/07 07:56

_id	cdrf2021_129
id	cdrf2021_129
authors	Fuyuan Liu, Min Chen, Lizhe Wang, Xiang Wang, and Cheng-Hung Lo
year	2021
title	Custom-Fit and Lightweight Optimization Design of Exoskeletons Using Parametric Conformal Lattice
source	Proceedings of the 2021 DigitalFUTURES The 3rd International Conference on Computational Design and Robotic Fabrication (CDRF 2021)
doi	https://doi.org/https://doi.org/10.1007/978-981-16-5983-6_12
summary	This paper presents an integrated design method for the customization and lightweight design of free-shaped wearable devices, illustrated by a lower limb exoskeleton. The customized design space is derived from the 3D scanning models. Based on the finite element analysis, the structural framework is determined through topology optimization with allowable strength. By means of generative design, the lattice library is constructed to fill the frames under different conformal algorithms. Finally, the proposed method is illustrated by the exoskeleton design case.
series	cdrf
email
last changed	2022/09/29 07:53

For more results click below: