cd /news/machine-learning/where-to-predict-in-retinal-oct-anat… · home › topics › machine-learning › article
[ARTICLE · art-170302] src=arxiv.org ↗ pub= topic=machine-learning verified=true sentiment=· neutral

Where to Predict in Retinal OCT? Anatomy-Guided Target Selection for I-JEPA

A study posted to arXiv (2611.17586v1) found that rectangular prediction targets guided by image intensity produced the highest linear-probe AUC when adapting the Image-based Joint-Embedding Predictive Architecture (I-JEPA) to retinal OCT scans, evaluated on glaucoma classification with the FairVision dataset. The researchers compared anatomy-guided rectangular targets, anatomy-shaped targets and coverage-constrained placement against I-JEPA's uniform masking while keeping the architecture and latent prediction objective unchanged. They concluded that anatomically meaningful target selection is promising for retinal OCT representation learning, but that greater anatomical specificity did not consistently improve performance.

by read1 min views1 publishedNov 16, 2026

arXiv:2611.17586v1 Announce Type: new Abstract: Self-supervised learning aims to learn meaningful representations from unlabeled data, potentially allowing models to capture subtle anatomical variations across patients and reduce the amount of labeled data needed. Image-based Joint-Embedding Predictive Architecture (I-JEPA) learns by predicting target-region representations from visible context rather than reconstructing pixels. However, its uniform target placement does not account for the anatomical structure of retinal OCT, where retinal tissue occupies only part of each B-scan. We investigate anatomy-guided target selection using image intensity and predicted retinal segmentation while retaining the I-JEPA architecture and latent prediction objective. We compare anatomy-guided rectangular targets, anatomy-shaped targets and coverage-constrained placement with uniform masking, and evaluate the learned representations using linear probing for glaucoma classification on FairVision. We found that rectangular targets guided by image intensity achieved the highest linear-probe AUC in our comparisons. We show that anatomically meaningful target selection is promising for retinal OCT representation learning, while greater anatomical specificity did not consistently improve performance.

── more in #machine-learning 4 stories · sorted by recency
── more on @i-jepa 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/where-to-predict-in-…] indexed:0 read:1min 2026-11-16 · —