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/

Zendy Plus

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

Zendy Tools

Zendy Open

Approaches for single-view reconstruction typically rely on viewpointannotations, silhouettes, the absence of background, multiple views of the sameinstance, a template shape, or symmetry. We avoid all such supervision andassumptions by explicitly leveraging the consistency between images ofdifferent object instances. As a result, our method can learn from largecollections of unlabelled images depicting the same object category. Our maincontributions are two ways for leveraging cross-instance consistency: (i)progressive conditioning, a training strategy to gradually specialize the modelfrom category to instances in a curriculum learning fashion; and (ii) neighborreconstruction, a loss enforcing consistency between instances having similarshape or texture. Also critical to the success of our method are: ourstructured autoencoding architecture decomposing an image into explicit shape,texture, pose, and background; an adapted formulation of differentialrendering; and a new optimization scheme alternating between 3D and poselearning. We compare our approach, UNICORN, both on the diverse syntheticShapeNet dataset - the classical benchmark for methods requiring multiple viewsas supervision - and on standard real-image benchmarks (Pascal3D+ Car, CUB) forwhich most methods require known templates and silhouette annotations. We alsoshowcase applicability to more challenging real-world collections (CompCars,LSUN), where silhouettes are not available and images are not cropped aroundthe object.

Share With Thy Neighbors: Single-View Reconstruction by Cross-Instance Consistency