Saved in:
Bibliographic Details
Main Authors: Cui, Zhe, Zhang, Shuxian, Lou, Kangzhi, Tran, Le-Nam
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.20193
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908342696280064
author Cui, Zhe
Zhang, Shuxian
Lou, Kangzhi
Tran, Le-Nam
author_facet Cui, Zhe
Zhang, Shuxian
Lou, Kangzhi
Tran, Le-Nam
contents This paper presents ProFi-Net, a novel few-shot learning framework for WiFi-based gesture recognition that overcomes the challenges of limited training data and sparse feature representations. ProFi-Net employs a prototype-based metric learning architecture enhanced with a feature-level attention mechanism, which dynamically refines the Euclidean distance by emphasizing the most discriminative feature dimensions. Additionally, our approach introduces a curriculum-inspired data augmentation strategy exclusively on the query set. By progressively incorporating Gaussian noise of increasing magnitude, the model is exposed to a broader range of challenging variations, thereby improving its generalization and robustness to overfitting. Extensive experiments conducted across diverse real-world environments demonstrate that ProFi-Net significantly outperforms conventional prototype networks and other state-of-the-art few-shot learning methods in terms of classification accuracy and training efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2504_20193
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ProFi-Net: Prototype-based Feature Attention with Curriculum Augmentation for WiFi-based Gesture Recognition
Cui, Zhe
Zhang, Shuxian
Lou, Kangzhi
Tran, Le-Nam
Machine Learning
This paper presents ProFi-Net, a novel few-shot learning framework for WiFi-based gesture recognition that overcomes the challenges of limited training data and sparse feature representations. ProFi-Net employs a prototype-based metric learning architecture enhanced with a feature-level attention mechanism, which dynamically refines the Euclidean distance by emphasizing the most discriminative feature dimensions. Additionally, our approach introduces a curriculum-inspired data augmentation strategy exclusively on the query set. By progressively incorporating Gaussian noise of increasing magnitude, the model is exposed to a broader range of challenging variations, thereby improving its generalization and robustness to overfitting. Extensive experiments conducted across diverse real-world environments demonstrate that ProFi-Net significantly outperforms conventional prototype networks and other state-of-the-art few-shot learning methods in terms of classification accuracy and training efficiency.
title ProFi-Net: Prototype-based Feature Attention with Curriculum Augmentation for WiFi-based Gesture Recognition
topic Machine Learning
url https://arxiv.org/abs/2504.20193