Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Methods for determining focal point and delay for ultrasound arrays and multichannel electronics

Rosenlind, Svante LU (2023) BMEM01 20231
Department of Biomedical Engineering
Abstract
High intensity focused ultrasound is a growing technique for tissue ablation, among
other uses, and given its destructive capabilities, there is a need for control of where the energy is delivered. There exist a number of methods for focusing such ultrasound arrays, but these often assume prior knowledge of the impulse response, or require extensive full-system simulations. This thesis explores schemes for calculating impulse and frequency response of simple but still nonhomogeneous media, and implements different focusing methods, the spatiotemporal inverse filter, the Gerchberg-Saxton algorithm, and gradient descent, to test them.

With a 128-channel transducer operating at 5 MHz, these techniques are carried
out in a simulated 2D... (More)
High intensity focused ultrasound is a growing technique for tissue ablation, among
other uses, and given its destructive capabilities, there is a need for control of where the energy is delivered. There exist a number of methods for focusing such ultrasound arrays, but these often assume prior knowledge of the impulse response, or require extensive full-system simulations. This thesis explores schemes for calculating impulse and frequency response of simple but still nonhomogeneous media, and implements different focusing methods, the spatiotemporal inverse filter, the Gerchberg-Saxton algorithm, and gradient descent, to test them.

With a 128-channel transducer operating at 5 MHz, these techniques are carried
out in a simulated 2D setting on water and concrete with first a straight edge and then an oblique one between the two media. With a focus depth of 5 cm, the techniques are able to clearly outperform the uncompensated results, and were able to produce feasible foci even for offset or multiple simultaneous foci locations.

Although the optimization-based method did fail to produce adequate results for
parts of the test, the overall investigation was seen as a successful venture, and that extension of the techniques to more complex media and 3D settings would be needed before any practical value can be realized. (Less)
Popular Abstract
Focusing methods for multi-channel ultrasound arrays show promise in theoretical investigation

When bats use ultrasound, they get information about objects in the space around
them, such as how far away they are, and in what direction they are. In modern hospitals however, the amount of information we receive from an ultrasound examination is much larger. That is mainly because we use more channels than bats. They have 2 ears and 1 mouth, and if the returning sound hits the right ear first, they can deduce that the object they are detecting is to the right. Medical ultrasound transducers commonly have 128 channels, and with the help of statistical processing, this results in a much clearer image.

Ultrasound is however not only used... (More)
Focusing methods for multi-channel ultrasound arrays show promise in theoretical investigation

When bats use ultrasound, they get information about objects in the space around
them, such as how far away they are, and in what direction they are. In modern hospitals however, the amount of information we receive from an ultrasound examination is much larger. That is mainly because we use more channels than bats. They have 2 ears and 1 mouth, and if the returning sound hits the right ear first, they can deduce that the object they are detecting is to the right. Medical ultrasound transducers commonly have 128 channels, and with the help of statistical processing, this results in a much clearer image.

Ultrasound is however not only used in examinations, but also to physically affect tissue. Kidney stones can be shattered, and tumours can be burned away by the friction caused by the ultrasound vibrations. For this to be safe, it needs to be focused properly, and that is where this thesis finds its objective. The human body is also full of different organs, and when passing from one to another, the sound is scattered, which adds additional difficulty.

The main modeling technique used was to view the sound waves as particles em-
anating spherically from each source. Then, it is relatively straightforward to simulate them bouncing around and spreading until they hit their target. Minor variations of this method were tried and compared. Once we know how each source can affect the target, the next step is to figure out which combination of inputs in which senders will result in the wanted output.

Here, as well, there are different methods, based on if you’re considering perfect
harmonic waves, or arbitrary inputs. The simplest one, based on classical optimization algorithms, is to start with a random selection of these perfect waves, with random amplitudes and relative phase differences. Then, you change it slightly in different directions, perhaps by changing one phase or one amplitude, and see if there is any improvement. Then you change the input in the direction that caused the largest improvement, and repeat until a satisfactory result is reached.

As the interest was mainly theoretical, simplifications were made to the geometry,
reducing the number of dimensions to two, and only considering two different types
of medium, with the border between them being straight. Because of this, no direct
conclusions about the medical viability could be drawn, but the methods produced
promising results, and the general take-away from the thesis is that this warrants further investigation, and the modelling techniques used here could potentially be useful in the future. (Less)
Please use this url to cite or link to this publication:
author
Rosenlind, Svante LU
supervisor
organization
alternative title
Metoder för bestämning av fokuspunkt och fördröjning för flerkanaliga ultraljudssystem
course
BMEM01 20231
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Focusing algorithms, k-wave, Spatiotemporal inverse filter, Gerchberg- Saxton algorithm, Focused ultrasound
language
English
additional info
2023-05
id
9114557
date added to LUP
2023-05-16 15:32:38
date last changed
2023-05-16 15:32:38
@misc{9114557,
  abstract     = {{High intensity focused ultrasound is a growing technique for tissue ablation, among
other uses, and given its destructive capabilities, there is a need for control of where the energy is delivered. There exist a number of methods for focusing such ultrasound arrays, but these often assume prior knowledge of the impulse response, or require extensive full-system simulations. This thesis explores schemes for calculating impulse and frequency response of simple but still nonhomogeneous media, and implements different focusing methods, the spatiotemporal inverse filter, the Gerchberg-Saxton algorithm, and gradient descent, to test them.

With a 128-channel transducer operating at 5 MHz, these techniques are carried
out in a simulated 2D setting on water and concrete with first a straight edge and then an oblique one between the two media. With a focus depth of 5 cm, the techniques are able to clearly outperform the uncompensated results, and were able to produce feasible foci even for offset or multiple simultaneous foci locations. 

Although the optimization-based method did fail to produce adequate results for
parts of the test, the overall investigation was seen as a successful venture, and that extension of the techniques to more complex media and 3D settings would be needed before any practical value can be realized.}},
  author       = {{Rosenlind, Svante}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Methods for determining focal point and delay for ultrasound arrays and multichannel electronics}},
  year         = {{2023}},
}