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

In recent years, the results of view-based 3D shape recognition methods havesaturated, and models with excellent performance cannot be deployed onmemory-limited devices due to their huge size of parameters. To address thisproblem, we introduce a compression method based on knowledge distillation forthis field, which largely reduces the number of parameters while preservingmodel performance as much as possible. Specifically, to enhance thecapabilities of smaller models, we design a high-performing large model calledGroup Multi-view Vision Transformer (GMViT). In GMViT, the view-level ViT firstestablishes relationships between view-level features. Additionally, to capturedeeper features, we employ the grouping module to enhance view-level featuresinto group-level features. Finally, the group-level ViT aggregates group-levelfeatures into complete, well-formed 3D shape descriptors. Notably, in bothViTs, we introduce spatial encoding of camera coordinates as innovativeposition embeddings. Furthermore, we propose two compressed versions based onGMViT, namely GMViT-simple and GMViT-mini. To enhance the trainingeffectiveness of the small models, we introduce a knowledge distillation methodthroughout the GMViT process, where the key outputs of each GMViT componentserve as distillation targets. Extensive experiments demonstrate the efficacyof the proposed method. The large model GMViT achieves excellent 3Dclassification and retrieval results on the benchmark datasets ModelNet,ShapeNetCore55, and MCB. The smaller models, GMViT-simple and GMViT-mini,reduce the parameter size by 8 and 17.6 times, respectively, and improve shaperecognition speed by 1.5 times on average, while preserving at least 90% of theclassification and retrieval performance.

Group Multi-View Transformer for 3D Shape Analysis with Spatial Encoding