BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20250110T023309Z
LOCATION:Hall B5 (2)\, B Block\, Level 5
DTSTART;TZID=Asia/Tokyo:20241203T135600
DTEND;TZID=Asia/Tokyo:20241203T141000
UID:siggraphasia_SIGGRAPH Asia 2024_sess104_papers_905@linklings.com
SUMMARY:MVImgNet2.0: A Larger-scale Dataset of Multi-view Images
DESCRIPTION:Technical Papers\n\nXiaoguang Han (SSE, The Chinese University
  of Hong Kong, Shenzhen; FNii, The Chinese University of Hong Kong, Shenzh
 en); Yushuang Wu, Luyue Shi, Haolin Liu, Hongjie Liao, and Lingteng Qiu (F
 Nii, The Chinese University of Hong Kong, Shenzhen; SSE, The Chinese Unive
 rsity of Hong Kong, Shenzhen); Weihao Yuan, Xiaodong Gu, and Zilong Dong (
 Alibaba); and Shuguang Cui (SSE, The Chinese University of Hong Kong, Shen
 zhen; FNii, The Chinese University of Hong Kong, Shenzhen)\n\nMVImgNet is 
 a large-scale dataset that contains multi-view images of ~220k real-world 
 objects in 238 classes. As a counterpart of ImageNet, it introduces 3D vis
 ual signals via multi-view shooting, making a soft bridge between 2D and 3
 D vision. This paper constructs the MVImgNet2.0 dataset that expands MVImg
 Net into a total of ~520k objects and 515 categories, which derives a 3D d
 ataset with a larger scale that is more comparable to ones in the 2D domai
 n. In addition to the expanded dataset scale and category range, MVImgNet2
 .0 is of a higher quality than MVImgNet owing to four new features: (i) mo
 st shoots capture 360-degree views of the objects, which can support the l
 earning of object reconstruction with completeness; (ii) the segmentation 
 manner is advanced to produce foreground object masks of higher accuracy; 
 (iii) a more powerful structure-from-motion method is adopted to derive th
 e camera pose for each frame of a lower estimation error; (iv) higher-qual
 ity dense point clouds are reconstructed via advanced methods for objects 
 captured in 360-degree views, which can serve for downstream applications.
  Extensive experiments confirm the value of the proposed MVImgNet2.0 in bo
 osting the performance of large 3D reconstruction models. MVImgNet2.0 will
  be public at luyues.github.io/mvimgnet2, including multi-view images of a
 ll 520k objects, the reconstructed high-quality point clouds, and data ann
 otation codes, hoping to inspire the broader vision community.\n\nRegistra
 tion Category: Full Access, Full Access Supporter\n\nLanguage Format: Engl
 ish Language\n\nSession Chair: Bernhard Kerbl (Technical University of Vie
 nna)
URL:https://asia.siggraph.org/2024/program/?id=papers_905&sess=sess104
END:VEVENT
END:VCALENDAR