Federated parameter-free DBSCAN clustering and its application in image recognition
Fang Cheng, Zilong Deng, Mustafa Muwafak Alobaedy, Xiaocun Huang · 2026
Read Federated parameter-free DBSCAN clustering and its application in image recognition by Fang Cheng, Zilong Deng, Mustafa Muwafak Alobaedy, Xiaocun Huang (2026) 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/plos-pone-0355161.txt.
Description
Fang Cheng and co-authors present in 2026 privacy preserving versions of a classic density based clustering method for data spread across many edge devices. Standard clustering needs central data, which breaks privacy, and its settings are hard to tune. Their two federated methods add noise for privacy at each client, find core points locally with an adaptive version, then merge those points on a server.
Eight real datasets, including a large handwriting image set, tested accuracy against standard and recent rivals. Both new methods clustered more accurately in their reports, and the stronger version also handled image grouping and segmentation well. The approach suits settings where raw data must stay local.
Catalog record
| CREATOR | Fang Cheng, Zilong Deng, Mustafa Muwafak Alobaedy, Xiaocun Huang |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37358 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |