Dendritic-Spine-Inspired Structural Optimization in a Lightweight Bowel Sound Analysis System for On-Device Colorectal Cancer Screening

View at publisher Summarise with VIRAT
Authors: Haojie Zhang, Enze Li, Yang Tan, Lin Shen, Kun Qian, Jing Li, Bin Hu, Björn Wolfgang Schuller, Yoshiharu Yamamoto
Cited by: 0
References: 14

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study presents a non-invasive AI-assisted Colorectal Cancer (CRC) screening system based on Bowel Sound (BS) analysis. A portable BS acquisition device is integrated with a dendritic-spine-inspired lightweight neural model that mimics biological synaptic pruning to remove redundant parameters while preserving key diagnostic features. The pruned model achieves high accuracy with minimal computational and memory costs, supporting on-device inference on microcontroller. Based on this model, we develop a complete device and system capable of efficient BS screening, demonstrating strong practicality for accessible CRC risk assessment.

Fields of Study & Subjects

Phonocardiography and Auscultation Techniques Colorectal Cancer Screening and Detection Artificial Intelligence in Healthcare

Publication details

Journal / source
International Conference on Consumer Electronics
Pages
1–4
Publication date
3 February 2026
Language
EN