arXiv subjects

Simon Bernard

Publications and source records attributed to Simon Bernard.

At least 19 recordsLinked to original sources

arXiv2026

Real-time tissue-equivalent measurement of individual clinical radiotherapy pulses

We apply the precision tools of cavity-enhanced absorption sensing to clinical oncology, demonstrating a dosimeter paradigm in which a centimeter-scale volume of water serves as a tissue-equivalent sensing medium. Our proof-of-concept, all-optical scheme achieves real-time readout of clinical radiation pulses with a nominal single-pulse resolution of 90 $\mu$Gy. This demonstration paves the way toward fiber-integrated, micron-scale devices for $\textit{in situ}$ universal absolute dosimetry during treatment.

physics.med-ph↗

arXiv2025

Predicting Patient Survival with Airway Biomarkers using nn-Unet/Radiomics

The primary objective of the AIIB 2023 competition is to evaluate the predictive significance of airway-related imaging biomarkers in determining the survival outcomes of patients with lung fibrosis.This study introduces a comprehensive three-stage approach. Initially, a segmentation network, namely nn-Unet, is employed to delineate the airway's structural boundaries. Subsequently, key features are extracted from the radiomic images centered around the trachea and an enclosing bounding box around the airway. This step is motivated by the potential presence of critical survival-related insights within the tracheal region as well as pertinent information encoded in the structure and dimensions of the airway. Lastly, radiomic features obtained from the segmented areas are integrated into an SVM classifier. We could obtain an overall-score of 0.8601 for the segmentation in Task 1 while 0.7346 for the classification in Task 2.

cs.CV↗

arXiv2024

TD-Paint: Faster Diffusion Inpainting Through Time Aware Pixel Conditioning

Diffusion models have emerged as highly effective techniques for inpainting, however, they remain constrained by slow sampling rates. While recent advances have enhanced generation quality, they have also increased sampling time, thereby limiting scalability in real-world applications. We investigate the generative sampling process of diffusion-based inpainting models and observe that these models make minimal use of the input condition during the initial sampling steps. As a result, the sampling trajectory deviates from the data manifold, requiring complex synchronization mechanisms to realign the generation process. To address this, we propose Time-aware Diffusion Paint (TD-Paint), a novel approach that adapts the diffusion process by modeling variable noise levels at the pixel level. This technique allows the model to efficiently use known pixel values from the start, guiding the generation process toward the target manifold. By embedding this information early in the diffusion process, TD-Paint significantly accelerates sampling without compromising image quality. Unlike conventional diffusion-based inpainting models, which require a dedicated architecture or an expensive generation loop, TD-Paint achieves faster sampling times without architectural modifications. Experimental results across three datasets show that TD-Paint outperforms state-of-the-art diffusion models while maintaining lower complexity.

cs.CV↗

arXiv2024

Radiation hardness of open Fabry-Perot microcavities

High-finesse microcavities offer a platform for compact, high-precision sensing by employing high-reflectivity, low-loss mirrors to create effective optical path lengths that are orders of magnitude larger than the device geometry. Here, we investigate the radiation hardness of Fabry-Perot microcavities formed from dielectric mirrors deposited on the tips of optical fibers. The microcavities are irradiated under both conventional (~0.1 Gy/s) and ultrahigh (FLASH, ~20 Gy/s) radiotherapy dose rates. Within our measurement sensitivity of ~40 ppm loss, we observe no degradation in the mirror absorption after irradiation with over 300 Gy accumulated dose. This result highlights the excellent radiation hardness of the dielectric mirrors forming the cavities, enabling new optics-based, real-time, in-vivo, tissue-equivalent radiation dosimeters with ~10 micron spatial resolution (our motivation), as well as other applications in high-radiation environments.

physics.med-ph↗

arXiv2024

An optically defined phononic crystal defect

We demonstrate a mechanical crystal with an optically programmable defect mode. By applying an optical spring to a single unit cell of a phononic crystal membrane, we smoothly transfer a single mechanical mode into the bandgap, thereby localizing its spatial profile from one spanning the entire crystal to one confined within a few unit cells. This localization is evidenced by an enhanced mechanical frequency shift commensurate with a 37-fold reduction in the mode's participating mass. Our results lay groundwork for a new class of optomechanical systems that control mechanical mode profile and participating mass.

physics.optics↗

arXiv2023

Random Forest Kernel for High-Dimension Low Sample Size Classification

High dimension, low sample size (HDLSS) problems are numerous among real-world applications of machine learning. From medical images to text processing, traditional machine learning algorithms are usually unsuccessful in learning the best possible concept from such data. In a previous work, we proposed a dissimilarity-based approach for multi-view classification, the Random Forest Dissimilarity (RFD), that perfoms state-of-the-art results for such problems. In this work, we transpose the core principle of this approach to solving HDLSS classification problems, by using the RF similarity measure as a learned precomputed SVM kernel (RFSVM). We show that such a learned similarity measure is particularly suited and accurate for this classification context. Experiments conducted on 40 public HDLSS classification datasets, supported by rigorous statistical analyses, show that the RFSVM method outperforms existing methods for the majority of HDLSS problems and remains at the same time very competitive for low or non-HDLSS problems.

stat.ML↗

arXiv2023

Multiple Noises in Diffusion Model for Semi-Supervised Multi-Domain Translation

In this work, we address the challenge of multi-domain translation, where the objective is to learn mappings between arbitrary configurations of domains within a defined set (such as $(D_1, D_2)\rightarrow{}D_3$, $D_2\rightarrow{}(D_1, D_3)$, $D_3\rightarrow{}D_1$, etc. for three domains) without the need for separate models for each specific translation configuration, enabling more efficient and flexible domain translation. We introduce Multi-Domain Diffusion (MDD), a method with dual purposes: i) reconstructing any missing views for new data objects, and ii) enabling learning in semi-supervised contexts with arbitrary supervision configurations. MDD achieves these objectives by exploiting the noise formulation of diffusion models, specifically modeling one noise level per domain. Similar to existing domain translation approaches, MDD learns the translation between any combination of domains. However, unlike prior work, our formulation inherently handles semi-supervised learning without modification by representing missing views as noise in the diffusion process. We evaluate our approach through domain translation experiments on BL3NDT, a multi-domain synthetic dataset designed for challenging semantic domain inversion, the BraTS2020 dataset, and the CelebAMask-HQ dataset.

cs.CL↗

arXiv2023

Towards Long-Term predictions of Turbulence using Neural Operators

This paper explores Neural Operators to predict turbulent flows, focusing on the Fourier Neural Operator (FNO) model. It aims to develop reduced-order/surrogate models for turbulent flow simulations using Machine Learning. Different model configurations are analyzed, with U-NET structures (UNO and U-FNET) performing better than the standard FNO in accuracy and stability. U-FNET excels in predicting turbulence at higher Reynolds numbers. Regularization terms, like gradient and stability losses, are essential for stable and accurate predictions. The study emphasizes the need for improved metrics for deep learning models in fluid flow prediction. Further research should focus on models handling complex flows and practical benchmarking metrics.

physics.flu-dyn↗

arXiv2022

Domain Translation via Latent Space Mapping

In this paper, we investigate the problem of multi-domain translation: given an element $a$ of domain $A$, we would like to generate a corresponding $b$ sample in another domain $B$, and vice versa. Acquiring supervision in multiple domains can be a tedious task, also we propose to learn this translation from one domain to another when supervision is available as a pair $(a,b)\sim A\times B$ and leveraging possible unpaired data when only $a\sim A$ or only $b\sim B$ is available. We introduce a new unified framework called Latent Space Mapping (\model) that exploits the manifold assumption in order to learn, from each domain, a latent space. Unlike existing approaches, we propose to further regularize each latent space using available domains by learning each dependency between pairs of domains. We evaluate our approach in three tasks performing i) synthetic dataset with image translation, ii) real-world task of semantic segmentation for medical images, and iii) real-world task of facial landmark detection.

cs.LG↗

arXiv2020

Random Forest for Dissimilarity-based Multi-view Learning

Many classification problems are naturally multi-view in the sense their data are described through multiple heterogeneous descriptions. For such tasks, dissimilarity strategies are effective ways to make the different descriptions comparable and to easily merge them, by (i) building intermediate dissimilarity representations for each view and (ii) fusing these representations by averaging the dissimilarities over the views. In this work, we show that the Random Forest proximity measure can be used to build the dissimilarity representations, since this measure reflects similarities between features but also class membership. We then propose a Dynamic View Selection method to better combine the view-specific dissimilarity representations. This allows to take a decision, on each instance to predict, with only the most relevant views for that instance. Experiments are conducted on several real-world multi-view datasets, and show that the Dynamic View Selection offers a significant improvement in performance compared to the simple average combination and two state-of-the-art static view combinations.

cs.LG↗

arXiv2020

A Novel Random Forest Dissimilarity Measure for Multi-View Learning

Multi-view learning is a learning task in which data is described by several concurrent representations. Its main challenge is most often to exploit the complementarities between these representations to help solve a classification/regression task. This is a challenge that can be met nowadays if there is a large amount of data available for learning. However, this is not necessarily true for all real-world problems, where data are sometimes scarce (e.g. problems related to the medical environment). In these situations, an effective strategy is to use intermediate representations based on the dissimilarities between instances. This work presents new ways of constructing these dissimilarity representations, learning them from data with Random Forest classifiers. More precisely, two methods are proposed, which modify the Random Forest proximity measure, to adapt it to the context of High Dimension Low Sample Size (HDLSS) multi-view classification problems. The second method, based on an Instance Hardness measurement, is significantly more accurate than other state-of-the-art measurements including the original RF Proximity measurement and the Large Margin Nearest Neighbor (LMNN) metric learning measurement.

cs.LG↗

arXiv2020

Monitored wet-etch removal of individual dielectric layers from high-finesse Bragg mirrors

It is prohibitively expensive to deposit customized dielectric coatings on individual optics. One solution is to batch-coat many optics with extra dielectric layers, then remove layers from individual optics as needed. Here we present a low-cost, single-step, monitored wet etch technique for reliably removing (or partially removing) individual SiO$_2$ and Ta$_2$O$_5$ dielectric layers, in this case from a high-reflectivity fiber mirror. By immersing in acid and monitoring off-band reflected light, we show it is straightforward to iteratively (or continuously) remove six bilayers. At each stage, we characterize the coating performance with a Fabry-P\'{e}rot cavity, observing the expected stepwise decrease in finesse from 92,000$\pm$3,000 to 3,950$\pm$50, finding no evidence of added optical losses. The etch also removes the fiber's sidewall coating after a single bilayer, and, after six bilayers, confines the remaining coating to a $\sim$50-$\mu$m-diameter pedestal at the center of the fiber tip. Vapor etching above the solution produces a tapered "pool cue" cladding profile, reducing the fiber diameter (nominally 125 $\mu$m) to $\sim$100 $\mu$m at an angle of $\sim$0.3$^\circ$ near the tip. Finally, we note that the data generated by this technique provides a sensitive estimate of the layers' optical depths. This technique could be readily adapted to free-space optics and other coatings.

physics.app-ph↗

arXiv2019

Flexure-Tuned Membrane-at-the-Edge Optomechanical System

We introduce a passively-aligned, flexure-tuned cavity optomechanical system in which a membrane is positioned microns from one end mirror of a Fabry-Perot optical cavity. By displacing the membrane through gentle flexure of its silicon supporting frame (i.e., to ~80 m radius of curvature (ROC)), we gain access to the full range of available optomechanical couplings, finding also that the optical spectrum exhibits none of the abrupt discontinuities normally found in "membrane-in-the-middle" (MIM) systems. More aggressive flexure (3 m ROC) enables >15 microns membrane travel, milliradian tilt tuning, and a wavelength-scale (1.64 $\pm$ 0.78 microns) membrane-mirror separation. We also provide a complete set of analytical expressions for this system's leading-order dispersive and dissipative optomechanical couplings. Notably, this system can potentially generate orders of magnitude larger linear dissipative or quadratic dispersive strong coupling parameters than is possible with a MIM system. Additionally, it can generate the same purely quadratic dispersive coupling as a MIM system, but with significantly suppressed linear dissipative back-action (and force noise).

quant-ph↗

arXiv2018

Swept-Frequency Drumhead Mechanical Resonators

We demonstrate a high-Q ($>5 \times 10^{6}$) swept-frequency membrane mechanical resonator achieving octave resonance tuning via an integrated heater and an unprecedented acceleration noise floor below 1 $μ$g Hz$^{-1/2}$ for frequencies above 50 kHz. This device is compatible with established batch fabrication techniques, and its optical readout is compatible with low-coherence light sources (e.g., a monochromatic light-emitting diode). The device can also be mechanically stabilized (or driven) with the same light source via bolometric optomechanics, and we demonstrate laser cooling from room temperature to 10 K. Finally, this method of frequency tuning is well-suited to fundamental studies of mechanical dissipation; in particular, we recover the dissipation spectra of many modes, identifying material damping and coupling to substrate resonances as the dominant loss mechanisms.

physics.app-ph↗

arXiv2018

Dynamic voting in multi-view learning for radiomics applications

Cancer diagnosis and treatment often require a personalized analysis for each patient nowadays, due to the heterogeneity among the different types of tumor and among patients. Radiomics is a recent medical imaging field that has shown during the past few years to be promising for achieving this personalization. However, a recent study shows that most of the state-of-the-art works in Radiomics fail to identify this problem as a multi-view learning task and that multi-view learning techniques are generally more efficient. In this work, we propose to further investigate the potential of one family of multi-view learning methods based on Multiple Classifiers Systems where one classifier is learnt on each view and all classifiers are combined afterwards. In particular, we propose a random forest based dynamic weighted voting scheme, which personalizes the combination of views for each new patient for classification tasks. The proposed method is validated on several real-world Radiomics problems.

cs.CV↗

arXiv2018

Ultrafast imaging of cell elasticity with optical microelastography

Elasticity is a fundamental cellular property that is related to the anatomy, functionality and pathological state of cells and tissues. However, current techniques based on cell deformation, atomic force microscopy or Brillouin scattering are rather slow and do not always accurately represent cell elasticity. Here, we have developed an alternative technique by applying shear wave elastography to the micrometer scale. Elastic waves were mechanically induced in live mammalian oocytes using a vibrating micropipette. These audible frequency waves were observed optically at 200,000 frames per second and tracked with an optical flow algorithm. Whole cell elasticity was then mapped using an elastography method inspired by the seismology field. Using this approach, we show that the elasticity of mouse oocyte is decreased when the oocyte cytoskeleton is disrupted with cytochalasin B. The technique is fast (less than 1 ms for data acquisition), precise (spatial resolution of a few micrometers), able to map internal cell structures, robust, and thus represents a tractable novel option for interrogating biomechanical properties of diverse cell types.

physics.med-ph↗

arXiv2018

Improve the performance of transfer learning without fine-tuning using dissimilarity-based multi-view learning for breast cancer histology images

Breast cancer is one of the most common types of cancer and leading cancer-related death causes for women. In the context of ICIAR 2018 Grand Challenge on Breast Cancer Histology Images, we compare one handcrafted feature extractor and five transfer learning feature extractors based on deep learning. We find out that the deep learning networks pretrained on ImageNet have better performance than the popular handcrafted features used for breast cancer histology images. The best feature extractor achieves an average accuracy of 79.30%. To improve the classification performance, a random forest dissimilarity based integration method is used to combine different feature groups together. When the five deep learning feature groups are combined, the average accuracy is improved to 82.90% (best accuracy 85.00%). When handcrafted features are combined with the five deep learning feature groups, the average accuracy is improved to 87.10% (best accuracy 93.00%).

cs.CV↗

arXiv2018

Dissimilarity-based representation for radiomics applications

Radiomics is a term which refers to the analysis of the large amount of quantitative tumor features extracted from medical images to find useful predictive, diagnostic or prognostic information. Many recent studies have proved that radiomics can offer a lot of useful information that physicians cannot extract from the medical images and can be associated with other information like gene or protein data. However, most of the classification studies in radiomics report the use of feature selection methods without identifying the machine learning challenges behind radiomics. In this paper, we first show that the radiomics problem should be viewed as an high dimensional, low sample size, multi view learning problem, then we compare different solutions proposed in multi view learning for classifying radiomics data. Our experiments, conducted on several real world multi view datasets, show that the intermediate integration methods work significantly better than filter and embedded feature selection methods commonly used in radiomics.

cs.CV↗