Saved in:
Bibliographic Details
Main Authors: Zhou, Chengwei, Majerus, Steve, Datta, Gourav
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.24178
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909813394374656
author Zhou, Chengwei
Majerus, Steve
Datta, Gourav
author_facet Zhou, Chengwei
Majerus, Steve
Datta, Gourav
contents Bladder pressure monitoring systems are increasingly vital in diagnosing and managing urinary tract dysfunction. Existing solutions rely heavily on hand-crafted features and shallow classifiers, limiting their adaptability to complex signal dynamics. We propose a one-layer streaming transformer model for real-time classification of bladder pressure states, operating on wavelet-transformed representations of raw time-series data. Our model incorporates temporal multi-head self-attention and state caching, enabling efficient online inference with high adaptability. Trained on a dataset of 91 patients with 20,000-80,000 samples each, our method demonstrates improved accuracy, higher energy- and latency-efficiency. Implementation considerations for edge deployment on low-power hardware, such as edge graphical processing units (GPU) and micro-controllers, are also discussed.
format Preprint
id arxiv_https___arxiv_org_abs_2509_24178
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BladderFormer: A Streaming Transformer for Real-Time Urological State Monitoring
Zhou, Chengwei
Majerus, Steve
Datta, Gourav
Signal Processing
Bladder pressure monitoring systems are increasingly vital in diagnosing and managing urinary tract dysfunction. Existing solutions rely heavily on hand-crafted features and shallow classifiers, limiting their adaptability to complex signal dynamics. We propose a one-layer streaming transformer model for real-time classification of bladder pressure states, operating on wavelet-transformed representations of raw time-series data. Our model incorporates temporal multi-head self-attention and state caching, enabling efficient online inference with high adaptability. Trained on a dataset of 91 patients with 20,000-80,000 samples each, our method demonstrates improved accuracy, higher energy- and latency-efficiency. Implementation considerations for edge deployment on low-power hardware, such as edge graphical processing units (GPU) and micro-controllers, are also discussed.
title BladderFormer: A Streaming Transformer for Real-Time Urological State Monitoring
topic Signal Processing
url https://arxiv.org/abs/2509.24178