arXiv subjects

Guillaume Lajoinie

Publications and source records attributed to Guillaume Lajoinie.

15 recordsLinked to original sources

arXiv2025

Characterization and virtualization of a medical ultrasound transducer

With the rise of data-driven ultrasound imaging technologies, realistic simulation of ultrasound fields and radio-frequency data is becoming increasingly important. Accurate transducer characterization is crucial for realistic simulations. In this work, we present a streamlined ultrasound characterization pipeline that creates a virtual transducer model from acoustic holography measurements of the ultrasound field. The pipeline enables the extraction of the transmit impulse response, the receive impulse response, the size of the transducer aperture, and the focal distance of the lens. The method relies on acoustic field projections using either the angular spectrum method or Rayleigh integrals. Additionally, we present a method to compensate for misalignment in the measurement setup, based on the angular spectrum of the holographic measurement. We demonstrate the application of the characterization pipeline to a P4-1 phased array transducer. The resulting virtual transducer model can be imported into the PROTEUS simulation software for the generation of physically realistic ultrasound fields and ultrasound radio-frequency data with arbitrary transmit settings. The characterization pipeline has been released as an open-source toolkit to enable the ultrasound community to perform transducer characterization with increased accuracy and efficiency.

physics.med-ph↗

arXiv2025

Steerable Anatomical Shape Synthesis with Implicit Neural Representations

Generative modeling of anatomical structures plays a crucial role in virtual imaging trials, which allow researchers to perform studies without the costs and constraints inherent to in vivo and phantom studies. For clinical relevance, generative models should allow targeted control to simulate specific patient populations rather than relying on purely random sampling. In this work, we propose a steerable generative model based on implicit neural representations. Implicit neural representations naturally support topology changes, making them well-suited for anatomical structures with varying topology, such as the thyroid. Our model learns a disentangled latent representation, enabling fine-grained control over shape variations. Evaluation includes reconstruction accuracy and anatomical plausibility. Our results demonstrate that the proposed model achieves high-quality shape generation while enabling targeted anatomical modifications.

cs.CV↗

arXiv2025

Waveform-Specific Performance of Deep Learning-Based Super-Resolution for Ultrasound Contrast Imaging

Resolving arterial flows is essential for understanding cardiovascular pathologies, improving diagnosis, and monitoring patient condition. Ultrasound contrast imaging uses microbubbles to enhance the scattering of the blood pool, allowing for real-time visualization of blood flow. Recent developments in vector flow imaging further expand the imaging capabilities of ultrasound by temporally resolving fast arterial flow. The next obstacle to overcome is the lack of spatial resolution. Super-resolved ultrasound images can be obtained by deconvolving radiofrequency (RF) signals before beamforming, breaking the link between resolution and pulse duration. Convolutional neural networks (CNNs) can be trained to locally estimate the deconvolution kernel and consequently super-localize the microbubbles directly within the RF signal. However, microbubble contrast is highly nonlinear, and the potential of CNNs in microbubble localization has not yet been fully exploited. Assessing deep learning-based deconvolution performance for non-trivial imaging pulses is therefore essential for successful translation to a practical setting, where the signal-to-noise ratio is limited, and transmission schemes should comply with safety guidelines. In this study, we train CNNs to deconvolve RF signals and localize the microbubbles driven by harmonic pulses, chirps, or delay-encoded pulse trains. Furthermore, we discuss potential hurdles for in-vitro and in-vivo super-resolution by presenting preliminary experimental results. We find that, whereas the CNNs can accurately localize microbubbles for all pulses, a short imaging pulse offers the best performance in noise-free conditions. However, chirps offer a comparable performance without noise, but are more robust to noise and outperform all other pulses in low-signal-to-noise ratio conditions.

physics.med-ph↗

arXiv2024

Vaporization dynamics of a super-heated water-in-oil droplet: modeling and numerical solution

The study of vapor bubble growth following droplet vaporization in a superheated liquid involves research areas such as hydrodynamics, heat transfer, mass transfer, and thermodynamics. The interplay between these multiscale aspects is strongly dependent on the geometry, the thermodynamic response, and the local physical properties of the system. To understand the role of each aspect of this complex mechanism we model super-heated droplet vaporization by coupling the equation of motion for bubble growth with the thermodynamics of phase change and heat transfer through the convection-diffusion equation. The semi-analytical model is validated with the analytical description for vapor bubble growth dominated either by inertia (Rayleigh) or by thermal diffusion (Plesset-Zwick), depending on droplet radius and degree of superheat. The effect of a mismatch of the thermal properties between the host liquid and the droplet is shown to be relevant only for low superheating, above which an increase in thermal diffusivity leads to a reduction in the rate of vaporization. At medium to high superheating, the droplet vaporizes completely without relying on thermal diffusion. At the point of complete vaporization, the potential energy within the system drives the bubble overshoots, which vary based on the droplet size and degree of superheat.

physics.flu-dyn↗

arXiv2023

A unifying Rayleigh-Plesset-type equation for bubbles in viscoelastic media

Understanding the ultrasound pressure-driven dynamics of microbubbles confined in viscoelastic materials is relevant for multiple biomedical applications, ranging from contrast-enhanced ultrasound imaging to ultrasound-assisted drug delivery. The volumetric oscillations of spherical bubbles is analyzed using the Rayleigh-Plesset equation, which describes the conservation of mass and momentum in the surrounding medium. Several studies have considered an extension of the Rayleigh-Plesset equation for bubbles embedded into viscoelastic media, but these are restricted to a particular choice of constitutive model and/or to small deformations. Here, we derive a unifying equation applicable to bubbles in viscoelastic media with arbitrary complex moduli and that can account for large bubble deformations. To derive this equation, we borrow concepts from finite-strain theory. We validate our approach by comparing the result of our model to previously published results and extend it to show how microbubbles behave in arbitrary viscoelastic materials. In particular, we use our viscoelastic Rayleigh-Plesset model to compute the bubble dynamics in benchmarked viscoelastic liquids and solids.

physics.flu-dyn↗

arXiv2022

Super-Resolved Microbubble Localization in Single-Channel Ultrasound RF Signals Using Deep Learning

Recently, super-resolution ultrasound imaging with ultrasound localization microscopy (ULM) has received much attention. However, ULM relies on low concentrations of microbubbles in the blood vessels, ultimately resulting in long acquisition times. Here, we present an alternative super-resolution approach, based on direct deconvolution of single-channel ultrasound radio-frequency (RF) signals with a one-dimensional dilated convolutional neural network (CNN). This work focuses on low-frequency ultrasound (1.7 MHz) for deep imaging (10 cm) of a dense cloud of monodisperse microbubbles (up to 1000 microbubbles in the measurement volume, corresponding to an average echo overlap of 94%). Data are generated with a simulator that uses a large range of acoustic pressures (5-250 kPa) and captures the full, nonlinear response of resonant, lipid-coated microbubbles. The network is trained with a novel dual-loss function, which features elements of both a classification loss and a regression loss and improves the detection-localization characteristics of the output. Whereas imposing a localization tolerance of 0 yields poor detection metrics, imposing a localization tolerance corresponding to 4% of the wavelength yields a precision and recall of both 0.90. Furthermore, the detection improves with increasing acoustic pressure and deteriorates with increasing microbubble density. The potential of the presented approach to super-resolution ultrasound imaging is demonstrated with a delay-and-sum reconstruction with deconvolved element data. The resulting image shows an order-of-magnitude gain in axial resolution compared to a delay-and-sum reconstruction with unprocessed element data.

physics.med-ph↗

arXiv2021

Time-resolved velocity and pressure field quantification in a flow-focusing device for ultrafast microbubble production

Flow-focusing devices have gained great interest in the past decade, due to their capability to produce monodisperse microbubbles for diagnostic and therapeutic medical ultrasound applications. However, up-scaling production to industrial scale requires a paradigm shift from single chip operation to highly parallelized systems. Parallelization gives rise to fluidic interactions between nozzles that, in turn, may lead to a decreased monodispersity. Here, we study the velocity and pressure field fluctuations in a single flow-focusing nozzle during bubble production. We experimentally quantify the velocity field inside the nozzle at 100 ns time resolution, and a numerical model provides insight into both the oscillatory velocity and pressure fields. Our results demonstrate that, at the length scale of the flow focusing channel, the velocity oscillations propagate at fluid dynamical time scale (order of microseconds) whereas the dominant pressure oscillations are linked to the bubble pinch-off and propagate at a much faster time scale (order of nanoseconds).

physics.flu-dyn↗

arXiv2021

Feedback-controlled microbubble generator producing one million monodisperse bubbles per second

Monodisperse lipid-coated microbubbles are a promising route to unlock the full potential of ultrasound contrast agents for medical diagnosis and therapy. Here, we present a stand-alone lab-on-a-chip instrument that allows microbubbles to be formed with high monodispersity at high production rates. Key to maintaining a long-term stable, controlled, and safe operation of the microfluidic device with full control over the output size distribution is an optical transmission based measurement technique that provides real-time information on production rate and bubble size. We feed the data into a feedback loop and demonstrate that this system can control the on-chip bubble radius (2.5 to 20 $μ$m) and the production rate up to $10^6$ bubbles/s. The freshly formed phospholipid-coated bubbles stabilize after their formation to a size approximately two times smaller than their initial on-chip bubble size without loss of monodispersity. The feedback control technique allows for full control over the size distribution of the agent and can aid the development of microfluidic platforms operated by non-specialist end users.

cond-mat.soft↗

arXiv2020

Ultrasound-enhanced mass transfer during single-bubble diffusive growth

Ultrasound is known to enhance surface bubble growth and removal in catalytic and microfluidic applications, yet the contributions of rectified diffusion and microstreaming phenomena towards mass transfer remain unclear. We quantify the effect of ultrasound on the diffusive growth of a single spherical CO$_2$ bubble growing on a substrate in supersaturated water. The time dependent bubble size, shape, oscillation amplitude and microstreaming flow field are resolved. We show and explain how ultrasound can enhance the diffusive growth of surface bubbles by up to two orders of magnitude during volumetric resonance. The proximity of the wall forces the bubble to oscillate non-spherically, thereby generating vigorous streaming during resonance that results in convection-dominated growth.

physics.flu-dyn↗

arXiv2020

Gas-Vapor Interplay in Plasmonic Bubble Shrinkage

The understanding of the shrinkage dynamics of plasmonic bubbles formed around metallic nanoparticles immersed in liquid and irradiated by a resonant light source is crucial for the usage of these bubbles in numerous applications. In this paper we experimentally show and theoretically explain that a plasmonic bubble during its shrinkage undergoes two different phases: first, a rapid partial bubble shrinkage governed by vapor condensation and, second, a slow diffusion-controlled bubble dissolution. The history of the bubble formation plays an important role in the shrinkage dynamics during the first phase, as it determines the gas-vapor ratio in the bubble composition. Higher laser powers lead to more vaporous bubbles, while longer pulses and higher dissolved air concentrations lead to more gaseous bubbles. The dynamics of the second phase barely depends on the history of bubble formation, i.e. laser power and pulse duration, but strongly on the dissolved air concentration, which defines the concentration gradient at the bubble interface. Finally, for the bubble dissolution in the second phase, with decreasing dissolved air concentration, we observe a gradual transition from a $R(t) \propto (t_0 - t) ^{1/3}$ scaling law to a $R(t) \propto (t_0 - t) ^{1/2}$ scaling law, where $t_0$ is the lifetime of the bubble and theoretically explain this transition.

physics.flu-dyn↗

arXiv2020

Plasmonic Bubble Nucleation and Growth in Water: Effect of Dissolved Air

Under continuous laser irradiation, noble metal nanoparticles immersed in water can quickly heat up, leading to the nucleation of so-called plasmonic bubbles. In this work, we want to further understand the bubble nucleation and growth mechanism. In particular, we quantitatively study the effect of the amount of dissolved air on the bubble nucleation and growth dynamics, both for the initial giant bubble, which forms shortly after switching on the laser and is mainly composed of vapor, and for the final life phase of the bubble, during which it mainly contains air expelled from water. We found that the bubble nucleation temperature depends on the gas concentration: the higher the gas concentration, the lower the bubble nucleation temperature. Also, the long-term diffusiondominated bubble growth is governed by the gas concentration. The radius of the bubbles grows as R(t)~t^1/3 for airequilibrated and air-oversaturated water. In contrast, in partially degassed water, the growth is much slower since, even for the highest temperature we achieve, the water remains undersaturated.

physics.flu-dyn↗

arXiv2019

Inverse Leidenfrost drop manipulation using menisci

Drops deposited on an evaporating liquid bath can be maintained in an inverse Leidenfrost state by the vapor emanating from the bath, making them levitate and hover almost without friction. These perfectly non-wetting droplets create a depression in the liquid interface that sustains their weight, which generates repellent forces when they approach a meniscus rising against a wall. Here, we study this reflection in detail, and show that frictionless Leidenfrost drops are a simple and efficient tool to probe the shape of an unknown interface. We then use the menisci to control the motion of the otherwise elusive drops. We create waveguides to direct and accelerate them and use parabolic walls to reflect and focus them. This could be particularly beneficial in the scale up of droplet cryopreservation processes: capillary interactions can be used to transport, gather and collect vitrified biological samples in absence of contact and contamination.

physics.flu-dyn↗

arXiv2019

Giant and explosive plasmonic bubbles by delayed nucleation

When illuminated by a laser, nano particles immersed in water can very quickly and strongly heat up, leading to the nucleation of so called vapor bubbles, which have huge application potential in e.g. solar light-harvesting, catalysis, and for medical applications.Whilst the long-time behavior of such bubbles has been well-studied, here, by employing ultra-high-speed imaging, we reveal the nucleation and early life phase of these bubbles. After some delay time after beginning of the illumination, a giant bubble explosively grows, up to a maximal radius of 80 um, and collapses again within 200 us (bubble life phase 1). The maximal bubble volume remarkably increases with decreasing laser power P. To explain this behavior, we measure the delay time from the beginning of the illumination up to nucleation,which drastically increases with decreasing laser power, leading to less total dumped energy . This dumped energy E shows a universal linear scaling relation. This finding supports that the initial giant bubble is a pure vapor bubble. In contrast, the delay time does depend on the gas concentration of the water, as gas pockets in the water facilitate an earlier vapor bubble nucleation, which leads to smaller delay times and lower bubble nucleation temperatures. After the collapse of the initial giant bubbles, first much smaller oscillating bubbles form out of the remaining gas nuclei (bubble life phase 2, up to typically 10 ms). Subsequently the known vaporization dominated growth phase takes over and the bubble stabilizes (life phase 3). In the final life phase 4 the bubble slowly grows by gas expelling due to heating of the surrounding. Our findings on the explosive growth and collapse during the early life phase of a vapor bubble have strong bearings on possible applications of such bubbles, affecting their risk assessment.

physics.app-ph↗

arXiv2019

Plasmonic Bubbles in n-Alkanes

In this paper we study the formation of microbubbles upon the irradiation of an array of plasmonic Au nanoparticles with a laser in n-alkanes ($C_{n}H_{2n+2}$, with n = 5-10). Two different phases in the evolution of the bubbles can be distinguished. In the first phase, which occurs after a delay time τd of about 100 μs, an explosive microbubble, reaching a diameter in the range from 10 μm to 100 μm, is formed. The exact size of this explosive microbubble barely depends on the carbon chain length of the alkane, but only on the laser power $P_l$. With increasing laser power, the delay time prior to bubble nucleation as well as the size of the microbubble both decrease. In the second phase, which sets in right after the collapse of the explosive microbubble, a new bubble forms and starts growing due to the vaporization of the surrounding liquid, which is highly gas rich. The final bubble size in this second phase strongly depends on the alkane chain length, namely it increases with decreasing number of carbon atoms. Our results have important implications for using plasmonic heating to control chemical reactions in organic solvents.

physics.chem-ph↗

arXiv2018

Monodisperse versus polydisperse ultrasound contrast agents: nonlinear response, sensitivity, and deep tissue imaging potential

Monodisperse microbubble ultrasound contrast agents have been proposed to further increase the signal-to-noise-ratio of contrast enhanced ultrasound imaging. Here, the sensitivity of a polydisperse preclinical agent was compared experimentally to that of its size- and acoustically-sorted derivatives by using narrowband pressure- and frequency-dependent scattering and attenuation measurements. The sorted monodisperse agents showed up to a two orders of magnitude increase in sensitivity, i.e. in the average scattering cross-section per bubble. Moreover, we demonstrate here, for the first time, that the highly nonlinear response of acoustically sorted microbubbles can be exploited to confine scattering and attenuation to the focal region of ultrasound fields used in clinical imaging. This property is a result of minimal prefocal scattering and attenuation and can be used to minimize shadowing effects in deep tissue imaging. Moreover, it potentially allows for more localized therapy using microbubbles through the spatial control of resonant microbubble oscillations.

physics.med-ph↗