Deep Convolutional State Space Model as Human Activity Recognizer

Wang Li

doi:10.1016/j.inffus.2025.103982

Deep Convolutional State Space Model as Human Activity Recognizer

2025-11-22

0

PRE

AI

DOI：10.1016/j.inffus.2025.103982

Original

Original request for help

Save

Abstract

En 中文

• This paper introduces a cutting-edge CNN-SSM model named DeepConvSSM. • DeepConvSSM is primarily built on the fusion of CNN with the state-space model. • Modern DeepConv Encoder (MDE) processes raw sensor inputs via CBRM layers. • Spatio-Temporal Mamba (STeM) applies spatial and temporal scan to refine embeddings. • It offers a computationally efficient and highly accurate activity recognizer.

Deep Convolutional State Space Model as Human Activity Recognizer

Abstract

Journal

Organization