AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

AFFMAE: Scalable Vision Pre-Training for High-Resolution Microscopy Segmentation on Desktop Hardware

David Smerkous, Zian Wang, Behzad Najafian

ECCV 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Low-rank compression and merging · 0 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
use AFFMAE instead of ViT-MAE for fine-tuninghigh-res segmentation · FPW (1024px)throughput img/s: 4 -> 20Improved
use AFFMAE instead of ViT-MAE for pre-trainingMAE pre-training · Pre-training 512pxthroughput img/s: 76 -> 151Improved
use AFFMAE instead of ViT-MAEMAE pre-training · Pre-training 512pxpeak memory GB: 29.7 -> 14.5Improved
use AFFMAE instead of ViT-MAEFPW segmentation · FPW (mIoU, 512/768/1024)mIoU: 0.606 -> 0.608 (512), 0.627 -> 0.622 (768), 0.630 -> 0.633 (1024)No clear effect
use AFFMAE instead of ViT-MAEFPW segmentation · FPW (1024px)Slits IoU: .494 -> .514Improved
in-domain EM pretraining instead of ImageNet-1K initMAE ViT · FPW 512pxmIoU: 0.494 -> 0.606Improved
add deep supervision (ablation)AFFMAE 23M · FPW ablation (d_s=0.4)mIoU: 0.5734 -> 0.5908Improved
reconstruct only from deepest stage instead of deep supervision (ablation)AFFMAE 23M · FPW ablationmIoU: 0.5908 -> 0.4353Hurt
increase token retention d_s from 0.25 to 0.5 (ablation)AFFMAE 23M · FPW ablationmIoU: 0.5729 -> 0.6009Improved
use d_s=0.4 rather than 0.5 (ablation)AFFMAE 23M · FPW ablationmIoU: 0.6009 -> 0.5908, GFLOPs 36 -> 30No clear effect
raise mask ratio from 0.50 to 0.75 (ablation)AFFMAE 23M · FPW ablationmIoU: 0.5908 -> 0.5739Hurt
Perlin masking instead of random masking (ablation)AFFMAE 23M · FPW ablation (d_s=0.5)mIoU: 0.5947 -> 0.6009Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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