PORTALE DELLA DIDATTICA

Ricerca CERCA
  KEYWORD

Biolab: Ingegneria Biomedica

Ultrasound imaging: beamforming and deep neural network

keywords BEAMFORMING, DEEP LEARNING, ULTRASOUND

Reference persons KRISTEN MARIKO MEIBURGER

Research Groups Biolab: Ingegneria Biomedica

Description Rationale
The process of forming a B-mode image relies on a specific technique called beamforming. Most ultrasound medical imaging system currently on the market implement standard Delay and Sum (DAS) beamforming. Recently, deep neural networks (DNNs) have been proposed for beamforming and segmentation of planewave ultrasound, but with some limitation (such as B-mode images include only circular, anechoic structures and a large dataset is used for the training.

Aim of the study:
The aim of this study is the investigation of the DNNsí generalizability for beamforming and segmenting structures of various shapes and echogenicity (Python preferred). Starting from Unet model trained with simulated data (Field II) the student will work to generalize the segmentation and formation of the B-mode images, for example investigating on the influence of the training using a dataset with an even higher variability of structures, as well as the inclusion of experimental images in the training dataset (using with Verasonics PolitoBIOLab).

See also  ultrasoundbeamformingcnn.pdf 


Deadline 01/04/2023      PROPONI LA TUA CANDIDATURA




© Politecnico di Torino
Corso Duca degli Abruzzi, 24 - 10129 Torino, ITALY
Contatti