3-D Instance Segmentation of MVS Buildings

Xu, Yanghui; Chen, Jiazhou; Lu, Shufang; Liang, Ronghua; Nan, Liangliang

doi:10.1109/tgrs.2022.3183567

Cited by 18 publications

(6 citation statements)

References 66 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These methods offer the potential to greatly enhance segmentation outcomes, particularly when they can leverage intrinsic texture information embedded within 3D mesh facets. This underscores the importance of staying abreast of developments in instance segmentation, exemplified by projects like Segment Anything [62], which push the boundaries of image segmentation models, and the Multi-View Stereo (MVS) building instance segmentation work [63], which demonstrates the possibilities of extracting 3D object instances in complex urban scenes. In addition, datasets like Ur-banBiS [64], which offer rich annotations and encompass extensive urban areas, represent essential resources for benchmarking and advancing segmentation techniques.…”

Section: Discussionmentioning

confidence: 99%

Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction

Leroux,

Germain,

Clabaut

et al. 2024

IJGI

View full text Add to dashboard Cite

Digital twins are increasingly gaining popularity as a method for simulating intricate natural and urban environments, with the precise segmentation of 3D objects playing an important role. This study focuses on developing a methodology for extracting buildings from textured 3D meshes, employing the PicassoNet-II semantic segmentation architecture. Additionally, we integrate Markov field-based contextual analysis for post-segmentation assessment and cluster analysis algorithms for building instantiation. Training a model to adapt to diverse datasets necessitates a substantial volume of annotated data, encompassing both real data from Quebec City, Canada, and simulated data from Evermotion and Unreal Engine. The experimental results indicate that incorporating simulated data improves segmentation accuracy, especially for under-represented features, and the DBSCAN algorithm proves effective in extracting isolated buildings. We further show that the model is highly sensible for the method of creating 3D meshes.

show abstract

Section: Discussionmentioning

confidence: 99%

Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction

Leroux,

Germain,

Clabaut

et al. 2024

IJGI

View full text Add to dashboard Cite

show abstract

“…Apart from the development of better models, several new EO-related vision datasets have been released over the years. Many of these datasets focus on various specific tasks, ranging from object detection [20], and different variants of segmentations [21] to change detection [22]. Apart from computer vision datasets of common scenes and terrestrial images, datasets like UAVid [23], and SpaceNet 4 [24] provide datasets with oblique or off-nadir imagery.…”

Section: Related Work and Motivationmentioning

confidence: 99%

UAVPal: A New Dataset for Semantic Segmentation in Complex Urban Landscape With Efficient Multiscale Segmentation

Maiti,

Elberink,

Vosselman

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Semantic segmentation has recently emerged as a prominent area of interest in Earth Observation. Several semantic segmentation datasets already exist, facilitating comparisons among different methods in complex urban scenes. However, most open high-resolution urban datasets are geographically skewed towards Europe and North America, while coverage of Southeast Asia is very limited. The considerable variation in city designs worldwide presents an obstacle to the applicability of computer vision models, especially when the training dataset lacks significant diversity. On the other hand, naively applying computationally expensive models leads to inefficacies and sometimes poor performance. To tackle the lack of data diversity, we introduce a new UAVPal dataset of complex urban scenes from the city of Bhopal, India. We complement this by introducing a novel dense predictor head and demonstrate that a well-designed head can efficiently take advantage of the multi-scale features to enhance the benefits of a strong feature extractor backbone. We design our segmentation head to learn the importance of features at various scales for each individual class and refine the final dense prediction accordingly. We tested our proposed head with a state-of-the-art backbone on multiple UAV datasets and a high-resolution satellite image dataset for land use land cover classification. We observed improved Intersection over Union (IoU) in various classes and up to 2% better mean IoU. Apart from the performance improvements, we also observed nearly 50% reduction in computing operations required when using the proposed head compared to the traditional segmentation head.

show abstract

“…Existing 3D instance segmentation methods can be categorized into distinct pipelines [15], [16]. Top-down approaches involve generating rough proposals such as bounding boxes or semantic segmentation results, followed by the refinement of these proposals [17].…”

Section: Introductionmentioning

confidence: 99%