WiFi CSI Based Temporal Activity Detection via Dual Pyramid Network

TL;DR

This paper introduces a novel Dual Pyramid Network for WiFi CSI-based temporal activity detection, combining semantic and local encoders with a new fusion mechanism, and demonstrates superior performance on a new dataset.

Contribution

The paper presents a new dual pyramid network architecture with innovative attention and fusion mechanisms for improved WiFi-based activity detection.

Findings

Outperforms existing baselines on the proposed dataset.

Effective separation of high and low-frequency features enhances detection.

New dataset with over 2,114 activity segments supports robust evaluation.

Abstract

We address the challenge of WiFi-based temporal activity detection and propose an efficient Dual Pyramid Network that integrates Temporal Signal Semantic Encoders and Local Sensitive Response Encoders. The Temporal Signal Semantic Encoder splits feature learning into high and low-frequency components, using a novel Signed Mask-Attention mechanism to emphasize important areas and downplay unimportant ones, with the features fused using ContraNorm. The Local Sensitive Response Encoder captures fluctuations without learning. These feature pyramids are then combined using a new cross-attention fusion mechanism. We also introduce a dataset with over 2,114 activity segments across 553 WiFi CSI samples, each lasting around 85 seconds. Extensive experiments show our method outperforms challenging baselines.