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

Single-view 3D shape retrieval is a challenging task that is increasinglyimportant with the growth of available 3D data. Prior work that has studiedthis task has not focused on evaluating how realistic occlusions impactperformance, and how shape retrieval methods generalize to scenarios whereeither the target 3D shape database contains unseen shapes, or the input imagecontains unseen objects. In this paper, we systematically evaluate single-view3D shape retrieval along three different axes: the presence of objectocclusions and truncations, generalization to unseen 3D shape data, andgeneralization to unseen objects in the input images. We standardize twoexisting datasets of real images and propose a dataset generation pipeline toproduce a synthetic dataset of scenes with multiple objects exhibitingrealistic occlusions. Our experiments show that training on occlusion-free dataas was commonly done in prior work leads to significant performance degradationfor inputs with occlusion. We find that that by first pretraining on oursynthetic dataset with occlusions and then finetuning on real data, we cansignificantly outperform models from prior work and demonstrate robustness toboth unseen 3D shapes and unseen objects.

Generalizing Single-View 3D Shape Retrieval to Occlusions and Unseen Objects