AI Paper Atlas

Research map / Detection and segmentation / CNN features

UPLiFT: Efficient Pixel-Dense Feature Upsampling with Local Attenders

Matthew Walmer, Saksham Suri, Anirud Aggarwal, Abhinav Shrivastava

CVPR 2026 · Detection and segmentation › CNN features · approach: Local features and context in CNNs · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
iterate LiFT to pixel-dense instead of single 2x step plus bilinearLiFT · COCO-Stuff segmentationmIoU: 58.28 -> 57.42Hurt
use UPLiFT (iterative, Local Attender) instead of LiFTDINOv2-S/14 · Cityscapes segmentationmIoU: 61.98 -> 65.38Improved
use UPLiFT instead of cross-attention AnyUpDINOv2-S/14 · ADE20K segmentationmIoU: 42.25 -> 42.97Improved
use UPLiFT instead of cross-attention AnyUpDINOv2-S/14 · Feature upsampling 448x448inference time: 146.7 -> 79.4 msImproved
use UPLiFT local-only attention instead of AnyUp for depthDINOv2-S/14 · COCO-Stuff depthdelta1: 61.32 vs 61.16Mixed
use local attention instead of quadratic cross-attention at ~1500 tokensUPLiFT · 624x624 imagesinference time: 2.5-5x fasterImproved
use UPLiFT instead of CFM for latent upscalingSD1.5 · COCO-5k 512->1024FID: 28.81 -> 24.23Improved
use UPLiFT instead of CFM for latent upscalingSD1.5 · FacesHQ super-resolutionFID: 1.36 -> 5.20Hurt
use UPLiFT instead of CFM for super-resolutionSD1.5 VAE · LHQ super-resolutionPSNR: 25.69 -> 26.70Improved
use larger UPLiFT model for VAE latentsUPLiFT · generative tasksfeature distributionImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy