Date of Publication :8th March 2025

Abstract: A portable system for rapid wheezing detection is proposed, designed for long-term patient monitoring and potential integration with other biomedical signal systems. The system utilizes a field-programmable gate array (FPGA) to accelerate the detection process. Sound signals are segmented into 2-second units and analyzed using a short-time Fourier transform to create spectrograms. These spectrograms are then processed using a series of techniques including 2D bilateral filtering, edge detection, multi -threshold image segmentation, morphological image processing, and image labeling. This processing extracts wheezing features based on computerized respiratory sound analysis (CORSA) standards. These features are used to train a support vector machine (SVM) fo r wheezing classification. The system is implemented on a Xilinx Virtex-6 FPGA ML605 platform. Testing demonstrates a wheezing recognition performance of 0.912, with the detection process operating at a clock frequency of 51.97 MHz, enabling rapid classification.

Reference :