Mitigating Resolution-Drift in Federated Learning: Case of Keypoint Detection
Taeheon Lim, Joohyung Lee, Kyungjae Lee, Jungchan Cho · 2025
Read Mitigating Resolution-Drift in Federated Learning: Case of Keypoint Detection by Taeheon Lim, Joohyung Lee, Kyungjae Lee, Jungchan Cho (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-23461.txt.
Description
Taeheon Lim, Joohyung Lee, Kyungjae Lee and Jungchan Cho name and tackle resolution-drift in federated learning, using human keypoint detection as the test case. Federated learning trains across scattered devices while data stay private, but past work stressed class mix and message cost. Here the trouble is that clients hold images at unlike resolutions, which drags down pose accuracy.
Their method, resolution-adaptive federated learning, uses heatmap-based knowledge distillation between high and low resolution outputs. That raises robustness to resolution change without overfitting, and the scheme drops into existing federated frameworks. Tests and theory both showed clear gains. A data plot analysis also showed why detail-keeping tasks differ from plain classification.
Catalog record
| CREATOR | Taeheon Lim, Joohyung Lee, Kyungjae Lee, Jungchan Cho |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 65 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-23461