GPU-Enabled Ultrasound Imaging Real-Time, Fully-Flexible Data - - PowerPoint PPT Presentation

▶
gpu enabled ultrasound imaging
SMART_READER_LITE
LIVE PREVIEW

GPU-Enabled Ultrasound Imaging Real-Time, Fully-Flexible Data - - PowerPoint PPT Presentation

Oct 29, 2022 498 likes •759 views

GPU-Enabled Ultrasound Imaging Real-Time, Fully-Flexible Data Processing Dr. Christoph Hennersperger Research Manager, Technical University of Munich Research Fellow, Trinity College Dublin CTO, OneProjects Manual Navigation Complex

slide-1
SLIDE 1
  • Dr. Christoph Hennersperger

Research Manager, Technical University of Munich Research Fellow, Trinity College Dublin CTO, OneProjects

GPU-Enabled Ultrasound Imaging

Real-Time, Fully-Flexible Data Processing

slide-2
SLIDE 2

1/27/2016 | Slide 2 GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Manual Navigation Reliability on Operator Complex Diagnostics

slide-3
SLIDE 3

1/27/2016 | Slide 3 GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Flexible Acquisition Guidance by/for Expert Intelligent Imaging

slide-4
SLIDE 4
  • Dr. Christoph Hennersperger

Research Manager, Technical University of Munich Research Fellow, Trinity College Dublin CTO, OneProjects

Real-Time, Fully-Flexible Data Acquisition Data Processing Toward Improved Diagnostics

slide-5
SLIDE 5

1/27/2016 | Slide 5

Brief Background on Ultrasound Imaging Workflow

Delay & Focus

Carrier signal Shape signal Pulse shape Active elements

Beamforming

  • Electronic delays for focusing
  • Applicable in transmit and

receive

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

B-mode image Scanline Scanline RF data Scanline envelope Scanline

Processing of received data

  • Radiofrequency data (RF)

scanline signals

  • Envelope detection

(Hilbert transform)

  • Subsampling (decimation)

Postprocessing for visualization

  • Subsequent filters pipeline

(e.g. speckle reduction)

  • Reduction of dynamic range

(Log-compression)

  • Transformation of scanlines to

image (Scan-conversion)

slide-6
SLIDE 6

1/27/2016 | Slide 6

The Need for a Software Defined Ultrasound Framework

Most ultrasound systems have limited flexibility

  • Implementation of major processing on DSPs, FPGAs or ASICs
  • Change of specific points require significant changes

Most ultrasound systems are closed systems

  • Access to images only through PACS
  • Proprietary access and interfaces

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Not usable for fast prototyping or research

slide-7
SLIDE 7

1/27/2016 | Slide 7

Software Defined Ultrasound Platform for Real-time Applications

Mission: Provide framework to allow covering research aspects from low-level US to high-level applications Key Design Properties

  • Data and module-driven approach
  • Fully software-defined platform (with GPU)

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Fully flexible design Fast prototyping Transparent storage Fully real-time

slide-8
SLIDE 8

1/27/2016 | Slide 8

General Ultrasound Processing Layout

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Transmit Beamforming (Aperture, Delays) Transmit & Receive Receive Beamforming (Delay and Sum with Apodization) Envelope detection incl. Frequency Compounding Log-Compression Scan-Conversion GPU accelerated Controllable via

slide-9
SLIDE 9

1/27/2016 | Slide 9

Beamforming on GPU

  • Parallelization for individual scanlines and samples
  • Aperture defines input data (number of channels)

Delay and sum beamformer in SUPRA Blocks operate on receive scanlines

  • Blocks process individual rx scanlines
  • Shared memory over local aperture (memory access)

Thread operate on receive samples

  • Individual threads process samples of scanlines
  • Local thread performs DAS over aperture (x,y) and depth (z) using

shared memory

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

+

τ2 τ1

τ3 τ4 τ5 τ6 τ7 τ8

slide-10
SLIDE 10

1/27/2016 | Slide 10

Imaging with SUPRA

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Fast Acquisitions Storage of Full Data Direct Configuration

slide-11
SLIDE 11

1/27/2016 | Slide 11

Qualitative Evaluation to Proprietary Scanline Imaging

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Cephasonics SUPRA Point Phantom Muscle Fibers Carotid Transverse Carotid Longitudinal

slide-12
SLIDE 12

1/27/2016 | Slide 12

From Scanline to Planewave Imaging

  • 64 scanlines
  • Max 300 Hz
  • 64 Angles
  • 300 Hz
  • 10 Angles
  • 1925 Hz

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Scanline Planewave Planewave

slide-13
SLIDE 13

1/27/2016 | Slide 13

Real-time Capabilities of Framework

  • Results for NVIDIA Jetson TX2, mobile GTX and GTX 1080

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

slide-14
SLIDE 14

1/27/2016 | Slide 14 GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Towards Improving Diagnostic Outcomes with US

slide-15
SLIDE 15

1/27/2016 | Slide 15

Ultrasound - Unique Abilities and Challenges

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Data interpretation

  • Image
  • Graph (network)
  • Continuous signals
slide-16
SLIDE 16

1/27/2016 | Slide 16

Ultrasound - Unique Abilities and Challenges

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Data interpretation

  • Image
  • Graph (network)
  • Continuous signals

Understanding of physics

  • Signals from reflection
  • Acoustic tissue properties
slide-17
SLIDE 17

1/27/2016 | Slide 17

Ultrasound - Unique Abilities and Challenges

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Data interpretation

  • Image
  • Graph (network)
  • Continuous signals

Understanding of physics

  • Signals from reflection
  • Acoustic tissue properties

Challenges and artefacts

  • Shadowing and enhancement
  • Nonlinearity of tissue propagation
  • Interference of waves
slide-18
SLIDE 18

1/27/2016 | Slide 19

Modeling Ultrasound as Arbitrarily Sampled Data

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

  • Overlapping US-slices
  • Resampling to regular grid
slide-19
SLIDE 19

1/27/2016 | Slide 20 GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

  • Overlapping US-slices
  • Resampling to regular grid
  • Process on regular graph

Loss of information regarding acquisition!

Modeling Ultrasound as Arbitrarily Sampled Data

slide-20
SLIDE 20

1/27/2016 | Slide 21

Modeling Ultrasound as Arbitrarily Sampled Data

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Better: Define graph on original samples

  • Graph nodes represent US samples
  • Graph edges represent spatial

structure Construct edges in “local” coordinates

slide-21
SLIDE 21

1/27/2016 | Slide 23

Modeling Ultrasound as Arbitrarily Sampled Data

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Better: Define graph on original samples

  • Graph nodes represent US samples
  • Graph edges represent spatial

structure Construct edges in “local” coordinates

1) Zu Berge, C. S., Declara, D., Hennersperger, C., Baust, M., & Navab, N. (2015, October). Real-time uncertainty visualization for B-mode ultrasound. IEEE SciVis 2015. 2) Hennersperger, C., Mateus, D., Baust, M., & Navab, N. (2014, September). A quadratic energy minimization framework for signal loss estimation from arbitrarily sampled ultrasound data, MICCAI 2015 3) Virga, S., Zettinig, O., Esposito, M., Pfister, K., Frisch, … & Hennersperger, C. (2016, October). Automatic force-compliant robotic ultrasound screening of abdominal aortic aneurysms, IEEE IROS 2016

slide-22
SLIDE 22

1/27/2016 | Slide 24

Modeling Ultrasound as Arbitrarily Sampled Data

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Better: Define graph on original samples

  • Graph nodes represent US samples
  • Graph edges represent spatial

structure Construct edges in “local” coordinates

1) Virga, S., Göbl, R., Baust, M., Navab, N., & Hennersperger, C. (2018). Use the force: deformation correction in robotic 3D ultrasound. International journal of computer assisted radiology and surgery, 1-9.

slide-23
SLIDE 23

1/27/2016 | Slide 25

Connecting the Dots

Improving Diagnostic Outcomes

  • Domain specific knowledge to improve

diagnostic benefit

  • Data science approaches with need for

data (end to end)

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

SUPRA as enabling technology

  • Fully open, flexible, and real-time
  • Tool for rapid demonstration and R&D
slide-24
SLIDE 24

1/27/2016 | Slide 26

Next: Try it Yourself!

GPU-Enabled Ultrasound Imaging | Christoph Hennersperger

Even without an ultrasound machine: https://github.com/IFL-CAMP/supra Rüdiger Göbl Christoph Hennersperger Nassir Navab

This project received funding from the European Union’s H2020 research and innovation programme (No 688279