000135569 001__ 135569
000135569 005__ 20240605121015.0
000135569 0247_ $$2doi$$a10.1109/TIM.2024.3392271
000135569 0248_ $$2sideral$$a138679
000135569 037__ $$aART-2024-138679
000135569 041__ $$aeng
000135569 100__ $$0(orcid)0000-0001-5709-1183$$aEnériz, Daniel$$uUniversidad de Zaragoza
000135569 245__ $$aLow-cost FPGA implementation of deep learning-based heart sound segmentation for real-time CVDs screening
000135569 260__ $$c2024
000135569 5060_ $$aAccess copy available to the general public$$fUnrestricted
000135569 5203_ $$aThe development of real-time, reliable, low-cost automatic phonocardiogram (PCG) analysis systems is critical for the early detection of cardiovascular diseases (CVDs), especially in countries with limited access to primary health care programs. Once the raw PCG acquired by the stethoscope has been preprocessed, the first key task is its segmentation into the fundamental heart sounds. For this purpose, an optimized hardware implementation of the segmentation algorithm is essential to attain a computer-aided diagnostic system based on PCGs. This article presents the optimization of a U-Net-based segmentation algorithm for its implementation in a low-end field-programmable gate array (FPGA) using low-resolution fixed-point data types. The optimization strategies seek to reduce the system latency while maintaining a constrained consumption of FPGA resources, aiming for a real-time response from the stethoscope data acquisition to the CVD detection. Experimental results prove a 64% decrease in latency compared to a baseline version, a 3.9% reduction of block random access memory (BRAM), which is the limiting resource of the design, and a 70% reduction in energy consumption. To the best of our knowledge, this is the first work to exhaustively study different optimization strategies for implementing a large 1-D U-Net-based model, achieving real-time fully characterized performance.
000135569 536__ $$9info:eu-repo/grantAgreement/ES/MICINN/PID2019-106570RB-I00$$9info:eu-repo/grantAgreement/ES/MICINN/PID2022-138785OB-I00
000135569 540__ $$9info:eu-repo/semantics/openAccess$$aby$$uhttp://creativecommons.org/licenses/by/3.0/es/
000135569 655_4 $$ainfo:eu-repo/semantics/article$$vinfo:eu-repo/semantics/publishedVersion
000135569 700__ $$aRodriguez-Almeida, Antonio J.
000135569 700__ $$aFabelo, Himar
000135569 700__ $$aOrtega, Samuel
000135569 700__ $$aBalea-Fernandez, Francisco J.
000135569 700__ $$aCallico, Gustavo M.
000135569 700__ $$0(orcid)0000-0002-5380-3013$$aMedrano, Nicolás$$uUniversidad de Zaragoza
000135569 700__ $$0(orcid)0000-0003-2361-1077$$aCalvo, Belén$$uUniversidad de Zaragoza
000135569 7102_ $$15008$$2250$$aUniversidad de Zaragoza$$bDpto. Ingeniería Electrón.Com.$$cÁrea Electrónica
000135569 773__ $$g73 (2024), 2003616 [16 pp.]$$pIEEE trans. instrum. meas.$$tIEEE Transactions on Instrumentation and Measurement$$x0018-9456
000135569 8564_ $$s2212622$$uhttps://zaguan.unizar.es/record/135569/files/texto_completo.pdf$$yVersión publicada
000135569 8564_ $$s3548679$$uhttps://zaguan.unizar.es/record/135569/files/texto_completo.jpg?subformat=icon$$xicon$$yVersión publicada
000135569 909CO $$ooai:zaguan.unizar.es:135569$$particulos$$pdriver
000135569 951__ $$a2024-06-05-10:50:43
000135569 980__ $$aARTICLE