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

Can we use sparse tokens for dense prediction, e.g., segmentation? Althoughtoken sparsification has been applied to Vision Transformers (ViT) toaccelerate classification, it is still unknown how to perform segmentation fromsparse tokens. To this end, we reformulate segmentation as a sparse encoding ->token completion -> dense decoding (SCD) pipeline. We first empirically showthat naively applying existing approaches from classification token pruning andmasked image modeling (MIM) leads to failure and inefficient training caused byinappropriate sampling algorithms and the low quality of the restored densefeatures. In this paper, we propose Soft-topK Token Pruning (STP) andMulti-layer Token Assembly (MTA) to address these problems. In sparse encoding,STP predicts token importance scores with a lightweight sub-network and samplesthe topK tokens. The intractable topK gradients are approximated through acontinuous perturbed score distribution. In token completion, MTA restores afull token sequence by assembling both sparse output tokens and prunedmulti-layer intermediate ones. The last dense decoding stage is compatible withexisting segmentation decoders, e.g., UNETR. Experiments show SCD pipelinesequipped with STP and MTA are much faster than baselines without token pruningin both training (up to 120% higher throughput and inference up to 60.6% higherthroughput) while maintaining segmentation quality.

Token Sparsification for Faster Medical Image Segmentation