SCUT-AutoALP: A Diverse Benchmark Dataset for Automatic Architectural Layout Parsing

Liu, Yubo; Lai, Yangting; Chen, Jianyong; Liang, Lingyu; Deng, Qiaoming

doi:10.1587/transinf.2020edl8076

Cited by 5 publications

(11 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We employ an FFN for inter-level connection classification. • Intra-level connection: Intra i ∈ {0, 1} 4 , defined in the same way as inter-level connections. Similar to inter-level connections, there are 16 connection types.…”

Section: Node Predicationmentioning

confidence: 99%

“…Previous works have either focused solely on floorplan semantic recognition, neglecting structural reconstruction [ZLYF19, LWG*21], or leaved structural reconstruction as a secondary stage [LWKF17, LZYZ21], relying heavily on heuristic post‐processing or additional optimization. In addition, the floor‐plan dataset also remains one of the main challenges, with existing publicly available datasets having either insufficient floor‐plan samples [LSK*15, LLC*20, LWKF17] or poor data quality [KYH*19], limiting the use and performance of deep neural networks. [LZYZ21] introduces a dataset comprising 7,000 annotated residential floorplans.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Raster‐to‐Graph: Floorplan Recognition via Autoregressive Graph Prediction with an Attention Transformer

Hu,

Wu,

et al. 2024

Computer Graphics Forum

View full text Add to dashboard Cite

Recognizing the detailed information embedded in rasterized floorplans is at the research forefront in the community of computer graphics and vision. With the advent of deep neural networks, automatic floorplan recognition has made tremendous breakthroughs. However, co‐recognizing both the structures and semantics of floorplans through one neural network remains a significant challenge. In this paper, we introduce a novel framework Raster‐to‐Graph, which automatically achieves structural and semantic recognition of floorplans. We represent vectorized floorplans as structural graphs embedded with floorplan semantics, thus transforming the floorplan recognition task into a structural graph prediction problem. We design an autoregressive prediction framework using the neural network architecture of the visual attention Transformer, iteratively predicting the wall junctions and wall segments of floorplans in the order of graph traversal. Additionally, we propose a large‐scale floorplan dataset containing over 10,000 real‐world residential floorplans. Our autoregressive framework can automatically recognize the structures and semantics of floorplans. Extensive experiments demonstrate the effectiveness of our framework, showing significant improvements on all metrics. Qualitative and quantitative evaluations indicate that our framework outperforms existing state‐of‐the‐art methods. Code and dataset for this paper are available at: https://github.com/HSZVIS/Raster-to-Graph.

show abstract

Section: Node Predicationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Raster‐to‐Graph: Floorplan Recognition via Autoregressive Graph Prediction with an Attention Transformer

Hu,

Wu,

et al. 2024

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…In both cases the drawing function outputs a drawing conditioned on the encoded input with some perturbation. Finally, an apartment layout drawing function was implemented based on the SCUT-ALP dataset (Liu et al, 2020) which uses Pix2Pix (Isola et al, 2018) to predict apartment layouts given a footprint.…”

Section: System Designmentioning

confidence: 99%

Desk Mate: A Collaborative Drawing Platform

Karsan

2023

Proceedings of the 28th Conference on Computer Aided Architectural Design Research in Asia (CAADRIA) [Volume 2]

View full text Add to dashboard Cite

Machine Learning (ML) in architecture is an emerging field with myriad potentials to impact the design process. Despite its many possibilities, ML is typically employed when the design problem is sufficiently defined, and further, is only integrated within software environments. Desk Mate is collaborative drawing machine that can be used early in the design process by coupling tangible tools like pens and trace paper with ML driven feedback and generation. Embedding physical tools that are familiar and intuitive with digital intelligence offers designers new ways of engaging with ML algorithms interactively, potentially changing the way the architectural industry approaches design problems. Desk Mate chains together image retrieval methods from machine vision with generative ML models like variational autoencoders (VAE) and generative adversarial networks (GANS) to react to design sketches as they are drawn. This pipeline allows Desk Mate to iterate through designs with the designer. Thus, Desk Mate demonstrates an interactive platform that collocates designer and machine as creative agents, facilitating drawing with ML driven feedback, potentially accelerating design iteration in the early stages of ideation.

show abstract

“…Constructing intelligent computer aided design (CAD) that achieves automatic layout design generation (LDG) with minimum manual adjustment is significant for architectural design. Recently, the problem of LDG has attracted evergrowing research interests in the field of computer vision and graphics [1]- [6] and these works have shown that automatic LDG is possible by learning from data.…”

Section: Introductionmentioning

confidence: 99%

“…The task of LDG aims to produce layout design can- subset II for urban plan in SCUT-AutoALP dataset [1], respectively; (c) shows samples for floor plan generation in RPLAN dataset [2].…”

Section: Introductionmentioning

confidence: 99%

Multiple Layout Design Generation via a GAN-Based Method with Conditional Convolution and Attention

ZHU,

LIU,

LIANG

et al. 2023

IEICE Trans. Inf. & Syst.

Self Cite

View full text Add to dashboard Cite

Recently, many AI-aided layout design systems are developed to reduce tedious manual intervention based on deep learning. However, most methods focus on a specific generation task. This paper explores a challenging problem to obtain multiple layout design generation (LDG), which generates floor plan or urban plan from a boundary input under a unified framework. One of the main challenges of multiple LDG is to obtain reasonable topological structures of layout generation with irregular boundaries and layout elements for different types of design. This paper formulates the multiple LDG task as an image-to-image translation problem, and proposes a conditional generative adversarial network (GAN), called LDGAN, with adaptive modules. The framework of LDGAN is based on a generator-discriminator architecture, where the generator is integrated with conditional convolution constrained by the boundary input and the attention module with channel and spatial features. Qualitative and quantitative experiments were conducted on the SCUT-AutoALP and RPLAN datasets, and the comparison with the state-of-the-art methods illustrate the effectiveness and superiority of the proposed LDGAN.

show abstract

SCUT-AutoALP: A Diverse Benchmark Dataset for Automatic Architectural Layout Parsing

Cited by 5 publications

References 13 publications

Raster‐to‐Graph: Floorplan Recognition via Autoregressive Graph Prediction with an Attention Transformer

Raster‐to‐Graph: Floorplan Recognition via Autoregressive Graph Prediction with an Attention Transformer

Desk Mate: A Collaborative Drawing Platform

Multiple Layout Design Generation via a GAN-Based Method with Conditional Convolution and Attention

Contact Info

Product

Resources

About