AI Paper Atlas

Research map / Federated learning and privacy / Personalized and heterogeneous FL

FeDepth: Federated Learning for Depth Estimation under Robot Heterogeneity

Ganghyeon Lee, inha Lee, Junhee Lee, Jeongeon Lee, Sung Whan Yoon, Kyungdon Joo

ECCV 2026 · Federated learning and privacy › Personalized and heterogeneous FL · approach: Federated prompt and LoRA tuning · 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 soft instead of hard clustering (ablation)FeDepth · M3ED HPE (NeWCRFs)Abs Rel: 0.264 -> 0.249Improved
raise cluster overlap from 0% to 20% (ablation)FeDepth · M3ED HPE (NeWCRFs)Abs Rel: 0.335 -> 0.249Improved
raise cluster overlap from 20% to 50% (ablation)FeDepth · M3ED HPE (NeWCRFs)Abs Rel: 0.249 -> 0.300Hurt
100% overlap (single cluster) (ablation)FeDepth · M3ED HPE (NeWCRFs)Abs Rel: 0.365 vs FedAvg 0.366Hurt
use descriptor-based clusters instead of ground-truth platform/environment label clusters (ablation)FeDepth · M3ED HPE (NeWCRFs)Abs Rel: 0.249 vs 0.471 (UGV only)Improved
FeDepth vs FedAvg on AdaBinsFeDepth · M3ED HPEAbs Rel: 0.434 -> 0.318Improved
FeDepth vs FedAvg on NeWCRFsFeDepth · M3ED HPEAbs Rel: 0.366 -> 0.249Improved
FeDepth vs FedAvg on DCDepthFeDepth · M3ED HPEAbs Rel: 0.351 -> 0.293Improved
FeDepth vs hard-clustering PACFL on BMR with AdaBinsFeDepth · BMRAbs Rel: 0.124 vs 0.126No clear effect
FeDepth vs PACFL on BMR with NeWCRFsFeDepth · BMRAbs Rel: 0.087 vs 0.089No clear effect
hard clustering PACFL with descriptor extractor on HPE AdaBinsPACFL · M3ED HPE (AdaBins)Abs Rel: 0.434 -> 0.602Hurt
pFedMe global model vs FeDepthpFedMe · M3ED HPE (NeWCRFs)Abs Rel: 0.563 vs 0.249Hurt

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