Adaptive prototype model for attribute-based multi-label few-shot action recognition

doi:10.1016/j.patcog.2026.114784

Adaptive prototype model for attribute-based multi-label few-shot action recognition

2026-09-16

0

PRE

AI

DOI：10.1016/j.patcog.2026.114784

Original

Original request for help

Save

Abstract

En 中文

• First multi-attribute few-shot action recognition framework (AMFAR). • TCM uses text features to build attribute-specific visual prototypes. • AAM assigns stochastic irrelevant attributes to prevent semantic drift. • Multi-Kinetics: first attribute-based multi-label action dataset. • AAPM achieves state-of-the-art on both multi-label and FSAR tasks.

Keywords:

Action recognition

Few-shot learning

Attribute-based multi-label recognition

Adaptive prototype model for attribute-based multi-label few-shot action recognition

Abstract

Journal

Organization

Cited Papers