Decoupled Prototype Matching with Vision Foundation Models for Few-Shot Industrial Object Detection
Hari Prasanth S. M, Nilusha Jayawickrama, Risto Ojala · 2026
Read Decoupled Prototype Matching with Vision Foundation Models for Few-Shot Industrial Object Detection by Hari Prasanth S. M, Nilusha Jayawickrama, Risto Ojala (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/arxiv-2604-26404.txt.
Description
Hari Prasanth S. M. and two colleagues tackled few-shot industrial object detection, where new items arrive with only a handful of labeled examples. Their framework builds class prototypes from few reference samples using vision foundation models, then matches query regions from a segmentation model against those prototypes. On three industrial datasets it beat the best training-free rivals by nearly 7 percent. New objects onboard from a few images, with no CAD models or large annotated sets needed.
Catalog record
| CREATOR | Hari Prasanth S. M, Nilusha Jayawickrama, Risto Ojala |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36301 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.26404