Augmenting Crowd-Sourced 3D Reconstructions Using Semantic Detections

Price, True; Schönberger, Johannes L.; Wei, Zhen; Pollefeys, Marc; Frahm, Jan‐Michael

doi:10.1109/cvpr.2018.00206

Cited by 9 publications

(4 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent works explicitly infer the light and reflectance properties of the objects in the scene from a set of unconstrained photo collections [16,26]. Others utilize semantic knowledge to reconstruct transient objects [23].…”

Section: Related Workmentioning

confidence: 99%

NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections

Martin-Brualla¹,

Radwan²,

Sajjadi³

et al. 2020

Preprint

View full text Add to dashboard Cite

We present a learning-based method for synthesizing novel views of complex outdoor scenes using only unstructured collections of in-the-wild photographs. We build on neural radiance fields (NeRF), which uses the weights of a multilayer perceptron to implicitly model the volumetric density and color of a scene. While NeRF works well on images of static subjects captured under controlled settings, it is incapable of modeling many ubiquitous, real-world phenomena in uncontrolled images, such as variable illumination or transient occluders. In this work, we introduce a series of extensions to NeRF to address these issues, thereby allowing for accurate reconstructions from unstructured image collections taken from the internet. We apply our system, which we dub NeRF-W, to internet photo collections of famous landmarks, thereby producing photorealistic, spatially consistent scene representations despite unknown and confounding factors, resulting in significant improvement over the state of the art.

show abstract

Section: Related Workmentioning

confidence: 99%

NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections

Martin-Brualla¹,

Radwan²,

Sajjadi³

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…Recent works have used a set of unconstrained photo collections to explicitly infer the light and reflectance of the objects in the scene [22,46]. Others make use of semantic information to restore transient objects [39].…”

Section: Related Workmentioning

confidence: 99%

Hallucinated Neural Radiance Fields in the Wild

Chen¹,

Zhang²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…As the photogrammetric model is not scaled, meaning the model does not have the real dimensions of the structure, a one-dimensional scale factor would be necessary. The latest crowdsourced reconstruction takes the advantage of object semantics, e.g., height distribution of people in the scene [41]. For heritage structure documentation, if the dimensions are necessary, archival data including at least one dimension or a tape measurement are suggested to be crowdsourced.…”

Section: Newcastle University Quadrangle Gateway Ukmentioning

confidence: 99%

Share Our Cultural Heritage (SOCH): Worldwide 3D Heritage Reconstruction and Visualization via Web and Mobile GIS

Dhonju

Xiao

Mills

et al. 2018

IJGI

View full text Add to dashboard Cite

Despite being of paramount importance to humanity, tangible cultural heritage is often at risk from natural and anthropogenic threats worldwide. As a result, heritage discovery and conservation remain a huge challenge for both developed and developing countries, with heritage sites often inadequately cared for, be it due to a lack of resources, nonrecognition of the value by local people or authorities, human conflict, or some other reason. This paper presents an online geo-crowdsourcing system, termed Share Our Cultural Heritage (SOCH), which can be utilized for large-scale heritage documentation and sharing. Supported by web and mobile GIS, cultural heritage data such as textual stories, locations, and images can be acquired via portable devices. These data are georeferenced and presented to the public via web-mapping. Using photogrammetric modelling, acquired images are used to reconstruct heritage structures or artefacts into 3D digital models, which are then visualized on the SOCH web interface to enable public interaction. This end-to-end system incubates an online virtual community to encourage public engagement, raise awareness, and stimulate cultural heritage ownership. It also provides valuable resources for cultural heritage exploitation, management, education, and monitoring over time.

show abstract

Augmenting Crowd-Sourced 3D Reconstructions Using Semantic Detections

Cited by 9 publications

References 48 publications

NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections

NeRF in the Wild: Neural Radiance Fields for Unconstrained Photo Collections

Hallucinated Neural Radiance Fields in the Wild

Share Our Cultural Heritage (SOCH): Worldwide 3D Heritage Reconstruction and Visualization via Web and Mobile GIS

Contact Info

Product

Resources

About