English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Reconstruction of a 3D shape from a single 2D image is a classical computervision problem, whose difficulty stems from the inherent ambiguity ofrecovering occluded or only partially observed surfaces. Recent methods addressthis challenge through the use of largely unstructured neural networks thateffectively distill conditional mapping and priors over 3D shape. In this work,we induce structure and geometric constraints by leveraging three coreobservations: (1) the surface of most everyday objects is often almost entirelyexposed from pairs of typical opposite views; (2) everyday objects oftenexhibit global reflective symmetries which can be accurately predicted fromsingle views; (3) opposite orthographic views of a 3D shape share consistentsilhouettes. Following these observations, we first predict orthographic 2.5Dvisible surface maps (depth, normal and silhouette) from perspective 2D images,and detect global reflective symmetries in this data; second, we predict theback facing depth and normal maps using as input the front maps and, whenavailable, the symmetric reflections of these maps; and finally, we reconstructa 3D mesh from the union of these maps using a surface reconstruction methodbest suited for this data. Our experiments demonstrate that our frameworkoutperforms state-of-the art approaches for 3D shape reconstructions from 2Dand 2.5D data in terms of input fidelity and details preservation.Specifically, we achieve 12% better performance on average in ShapeNetbenchmark dataset, and up to 19% for certain classes of objects (e.g., chairsand vessels).

Front2Back: Single View 3D Shape Reconstruction via Front to Back Prediction