Showing 1–9 of 9 results for author: Lacoste, B

A new dataset for measuring the performance of blood vessel segmentation methods under distribution shifts

Authors: Matheus Viana da Silva, Natália de Carvalho Santos, Julie Ouellette, Baptiste Lacoste, Cesar Henrique Comin

Abstract: Creating a dataset for training supervised machine learning algorithms can be a demanding task. This is especially true for medical image segmentation since one or more specialists are usually required for image annotation, and creating ground truth labels for just a single image can take up to several hours. In addition, it is paramount that the annotated samples represent well the different cond… ▽ More Creating a dataset for training supervised machine learning algorithms can be a demanding task. This is especially true for medical image segmentation since one or more specialists are usually required for image annotation, and creating ground truth labels for just a single image can take up to several hours. In addition, it is paramount that the annotated samples represent well the different conditions that might affect the imaged tissues as well as possible changes in the image acquisition process. This can only be achieved by considering samples that are typical in the dataset as well as atypical, or even outlier, samples. We introduce VessMAP, a heterogeneous blood vessel segmentation dataset acquired by carefully sampling relevant images from a larger non-annotated dataset. A methodology was developed to select both prototypical and atypical samples from the base dataset, thus defining an assorted set of images that can be used for measuring the performance of segmentation algorithms on samples that are highly distinct from each other. To demonstrate the potential of the new dataset, we show that the validation performance of a neural network changes significantly depending on the splits used for training the network. △ Less

Submitted 18 April, 2024; v1 submitted 11 January, 2023; originally announced January 2023.

Comments: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible

An Analysis of the Influence of Transfer Learning When Measuring the Tortuosity of Blood Vessels

Authors: Matheus V. da Silva, Julie Ouellette, Baptiste Lacoste, Cesar H. Comin

Abstract: Characterizing blood vessels in digital images is important for the diagnosis of many types of diseases as well as for assisting current researches regarding vascular systems. The automated analysis of blood vessels typically requires the identification, or segmentation, of the blood vessels in an image or a set of images, which is usually a challenging task. Convolutional Neural Networks (CNNs) h… ▽ More Characterizing blood vessels in digital images is important for the diagnosis of many types of diseases as well as for assisting current researches regarding vascular systems. The automated analysis of blood vessels typically requires the identification, or segmentation, of the blood vessels in an image or a set of images, which is usually a challenging task. Convolutional Neural Networks (CNNs) have been shown to provide excellent results regarding the segmentation of blood vessels. One important aspect of CNNs is that they can be trained on large amounts of data and then be made available, for instance, in image processing software for wide use. The pre-trained CNNs can then be easily applied in downstream blood vessel characterization tasks such as the calculation of the length, tortuosity, or caliber of the blood vessels. Yet, it is still unclear if pre-trained CNNs can provide robust, unbiased, results on downstream tasks when applied to datasets that they were not trained on. Here, we focus on measuring the tortuosity of blood vessels and investigate to which extent CNNs may provide biased tortuosity values even after fine-tuning the network to the new dataset under study. We show that the tortuosity values obtained by a CNN trained from scratch on a dataset may not agree with those obtained by a fine-tuned network that was pre-trained on a dataset having different tortuosity statistics. In addition, we show that the improvement in segmentation performance when fine-tuning the network does not necessarily lead to a respective improvement on the estimation of the tortuosity. To mitigate the aforementioned issues, we propose the application of specific data augmentation techniques even in situations where they do not improve segmentation performance. △ Less

Submitted 10 January, 2022; v1 submitted 19 November, 2021; originally announced November 2021.

Comments: Correction of typos. Change of mathematical notation. Addition of new sections, appendix, and supplementary material

Stabilization of phase noise in spin torque nano oscillators by a phase locked loop

Authors: Steffen Wittrock, Martin Kreißig, Bertrand Lacoste, Artem Litvinenko, Philippe Talatchian, Florian Protze, Frank Ellinger, Ricardo Ferreira, Romain Lebrun, Paolo Bortolotti, Liliana Buda-Prejbeanu, Ursula Ebels, Vincent Cros

Abstract: The main limitation in order to exploit spin torque nano-oscillators (STNOs) in various potential applications is their large phase noise. In this work, we demonstrate its efficient reduction by a highly reconfigurable, compact, specifically on-chip designed PLL based on custom integrated circuits. First, we thoroughly study the parameter space of the PLL+STNO system experimentally. Second, we pre… ▽ More The main limitation in order to exploit spin torque nano-oscillators (STNOs) in various potential applications is their large phase noise. In this work, we demonstrate its efficient reduction by a highly reconfigurable, compact, specifically on-chip designed PLL based on custom integrated circuits. First, we thoroughly study the parameter space of the PLL+STNO system experimentally. Second, we present a theory which describes the locking of a STNO to an external signal in a general sense. In our developed theory, we do not restrict ourselves to the case of a perfect phase locking but also consider phase slips and the corresponding low offset frequency $1/f^2$ noise, so far the main drawback in such systems. Combining experiment and theory allows us to reveal complex parameter dependences of the system's phase noise. The results provide an important step for the optimization of noise properties and thus leverage the exploitation of STNOs in prospective real applications. △ Less

Submitted 25 October, 2021; originally announced October 2021.

Magnetic oscillations Excited by Concurrent Spin Injection from a Tunneling Current and a Spin Hall Current

Authors: M. Tarequzzaman, T. Böhnert, M. Decker, J. D. Costa, J. Borme, B. Lacoste, E. Paz, A. S. Jenkins, S. Serrano-Guisan, C. H. Back, R. Ferreira, P. P. Freitas

Abstract: In this paper, a 3-terminal spin-transfer torque nano-oscillator (STNO) is studied using the concurrent spin injection of a spin-polarized tunneling current and a spin Hall current exciting the free layer into dynamic regimes beyond what is achieved by each individual mechanism. The pure spin injection is capable of inducing oscillations in the absence of charge currents effectively reducing the c… ▽ More In this paper, a 3-terminal spin-transfer torque nano-oscillator (STNO) is studied using the concurrent spin injection of a spin-polarized tunneling current and a spin Hall current exciting the free layer into dynamic regimes beyond what is achieved by each individual mechanism. The pure spin injection is capable of inducing oscillations in the absence of charge currents effectively reducing the critical tunneling current to zero. This reduction of the critical charge currents can improve the endurance of both STNOs and non-volatile magnetic memories (MRAM) devices. It is shown that the system response can be described in terms of an injected spin current density $J_s$ which results from the contribution of both spin injection mechanisms, with the tunneling current polarization $p$ and the spin Hall angle $θ$ acting as key parameters determining the efficiency of each injection mechanism. The experimental data exhibits an excellent agreement with this model which can be used to quantitatively predict the critical points ($J_s = -2.26\pm 0.09 \times 10^9 \hbar/e$ A/m$^2$) and the oscillation amplitude as a function of the input currents. In addition, the fitting of the data also allows an independent confirmation of the values estimated for the spin Hall angle and tunneling current polarization as well as the extraction of the dam** $α= 0.01$ and non-linear dam** $Q = 3.8\pm 0.3$ parameters. △ Less

Submitted 6 February, 2018; originally announced February 2018.

Comments: 8pages, 4 figures

Injection locking at 2f of spin torque oscillators under influence of thermal noise

Authors: M. Tortarolo, B. Lacoste, J. Hem, C. Dieudonné, M. -C. Cyrille, J. A. Katine, D. Mauri, A. Zeltser, L. D. Buda-Prejbeanu, U. Ebels

Abstract: Experiments, numerical simulations and an analytic model were developed to elucidate the effects of noise in the synchronized state of a tunnel junction based spin torque nano oscillator (STNO). It is demonstrated that in the in plane magnetized structure, while the frequency is locked, much higher reference currents are needed to reduce the noise by phase locking. Our analysis shows that it is po… ▽ More Experiments, numerical simulations and an analytic model were developed to elucidate the effects of noise in the synchronized state of a tunnel junction based spin torque nano oscillator (STNO). It is demonstrated that in the in plane magnetized structure, while the frequency is locked, much higher reference currents are needed to reduce the noise by phase locking. Our analysis shows that it is possible to control the phase noise by the reference microwave current (IRF) and that it can be further reduced by increasing the bias current (IDC) of the oscillator, kee** the reference current in feasible limits for applications. △ Less

Submitted 19 June, 2017; originally announced June 2017.

Synthetic ferrimagnet spin transfer torque oscillator: model and non-linear properties

Authors: Bertrand Lacoste, Miguel Romera, Ursula Ebels, Liliana Buda-Prejbeanu

Abstract: The non-linear parameters of spin-torque oscillators based on a synthetic ferrimagnet free layer (two coupled layers) are computed. The analytical expressions are compared to macrospin simulations in the case of a synthetic ferrimagnet excited by a current spin-polarized by an external fixed layer. It is shown that, of the two linear modes, acoustic and optical, only one is excited at a time, and… ▽ More The non-linear parameters of spin-torque oscillators based on a synthetic ferrimagnet free layer (two coupled layers) are computed. The analytical expressions are compared to macrospin simulations in the case of a synthetic ferrimagnet excited by a current spin-polarized by an external fixed layer. It is shown that, of the two linear modes, acoustic and optical, only one is excited at a time, and therefore the self-sustained oscillations are similar to the dynamics of a single layer. However, the non-linear parameters values can be controlled by the parameters of the synthetic ferrimagnet. With a strong coupling between the two layers and asymmetric layers (different thicknesses), it is demonstrated that the non-linear frequency shift can be reduced, which results in the reduction of the linewidth of the power spectral density. For a particular applied field, the non-linear parameter can even vanish; this corresponds to a transition between a red-shift and a blue-shift frequency dependence on the current and a linewidth reduction to the linear linewidth value. △ Less

Submitted 3 May, 2017; originally announced May 2017.

Spin torque driven dynamics of a coupled two layer structure: interplay between conservative and dissipative coupling

Authors: M. Romera, B. Lacoste, U. Ebels, L. D. Buda-Prejbeanu

Abstract: In this manuscript the general concepts of spin wave theory are adapted to the dynamics of a self-polarized system based on two layers coupled via interlayer exchange (conservative coupling) and mutual spin torque (dissipative coupling). An analytical description of the non-linear dynamics is proposed and validated through numerical simulations. In contrast to the single layer model, the phase equ… ▽ More In this manuscript the general concepts of spin wave theory are adapted to the dynamics of a self-polarized system based on two layers coupled via interlayer exchange (conservative coupling) and mutual spin torque (dissipative coupling). An analytical description of the non-linear dynamics is proposed and validated through numerical simulations. In contrast to the single layer model, the phase equation of the coupled system has a contribution coming from the dissipative part of the LLGS equation. It is shown that this is a major contribution to the frequency mandatory to describe well the most basic features of the dynamics of coupled systems. Using the proposed model a specific feature of coupled dynamics is addressed: the redshift to blueshift transition observed in the frequency current dependence of this kind of exchange coupled systems upon increasing the applied field. It is found that the blueshift regime can only occur in a region of field where the two linear eigenmodes contribute equally to the steady state mode (i.e. high mode hybridization). Finally, a general perturbed Hamiltonian equation for the coupled system is proposed. △ Less

Submitted 29 January, 2016; originally announced January 2016.

Comments: 16 pages, 7 figues

Journal ref: Phys. Rev. B 94, 094432 (2016)

Modulating spin transfer torque switching dynamics with two orthogonal spin-polarizers by varying the cell aspect ratio

Authors: B. Lacoste, M. Marins de Castro, T. Devolder, R. C. Sousa, L. D. Buda-Prejbeanu, S. Auffret, U. Ebels, C. Ducruet, I. L. Prejbeanu, L. Vila, B. Rodmacq, B. Dieny

Abstract: We study in-plane magnetic tunnel junctions with additional perpendicular polarizer for subnanosecond-current-induced switching memories. The spin-transfer-torque switching dynamics was studied as a function of the cell aspect ratio both experimentally and by numerical simulations using the macrospin model. We show that the anisotropy field plays a significant role in the dynamics, along with the… ▽ More We study in-plane magnetic tunnel junctions with additional perpendicular polarizer for subnanosecond-current-induced switching memories. The spin-transfer-torque switching dynamics was studied as a function of the cell aspect ratio both experimentally and by numerical simulations using the macrospin model. We show that the anisotropy field plays a significant role in the dynamics, along with the relative amplitude of the two spin-torque contributions. This was confirmed by micromagnetic simulations. Real-time measurements of the reversal were performed with samples of low and high aspect ratio. For low aspect ratios, a precessional motion of the magnetization was observed and the effect of temperature on the precession coherence was studied. For high aspect ratios, we observed magnetization reversals in less than 1 ns for high enough current densities, the final state being controlled by the current direction in the magnetic tunnel junction cell. △ Less

Submitted 3 December, 2014; v1 submitted 23 July, 2014; originally announced July 2014.

Comments: 6 pages, 7 figures

Journal ref: Phys. Rev. B 90, 224404 (2014)

Anisotropic magnetothermoelectric power of ferromagnetic thin films

Authors: M. S. Anwar, B. Lacoste, J. Aarts

Abstract: We compare the behavior of the magnetothermoelectric power (MTEP)in metallic ferromagnetic thin films of Ni80Fe20 (Permalloy; Py), Co and CrO2 at temperatures in the range of 100 K to 400 K. In 25 nm thick Py films and 50 nm thick Co films both the anisotropic magnetoresistance (AMR) and MTEP show a relative change in resistance and thermoelectric power (TEP) of the order of 0.2% when the magnetic… ▽ More We compare the behavior of the magnetothermoelectric power (MTEP)in metallic ferromagnetic thin films of Ni80Fe20 (Permalloy; Py), Co and CrO2 at temperatures in the range of 100 K to 400 K. In 25 nm thick Py films and 50 nm thick Co films both the anisotropic magnetoresistance (AMR) and MTEP show a relative change in resistance and thermoelectric power (TEP) of the order of 0.2% when the magnetic field is reversed, and in both cases there is no significant change in AMR or MTEP any more after the saturation field has been reached. Surprisingly, both Py and Co films have opposite MTEP behavior although both have the same sign for AMR and TEP. The data on 100 nm films of fully spin-polarized CrO2, grown both on TiO2 and on sapphire, show a different picture. The MTEP behavior at low fields shows peaks similar to the AMR in these films, with variations up to 1%. With increasing field both the MR and the MTEP variations keeps growing, with MTEP showing relative changes of 1.5% with the thermal gradient along the b-axis and even 20% with the gradient along the c-axis, with an intermediate value of 3% for the film on sapphire. It appears that the low-field effects are due to magnetic domain switching, while the high-field effects are intrinsic to the electronic structure of CrO2. △ Less

Submitted 15 June, 2017; v1 submitted 7 June, 2012; originally announced June 2012.

Comments: 6 pages, 7 figures