Bruno Galerne

Bruno Galerne, Lara Raad, José Lezama et al.

Neural style transfer (NST) is a deep learning technique that produces an unprecedentedly rich style transfer from a style image to a content image. It is particularly impressive when it comes to transferring style from a painting to an image. NST was originally achieved by solving an optimization problem to match the global statistics of the style image while preserving the local geometric features of the content image. The two main drawbacks of this original approach is that it is computationally expensive and that the resolution of the output images is limited by high GPU memory requirements. Many solutions have been proposed to both accelerate NST and produce images with larger size. However, our investigation shows that these accelerated methods all compromise the quality of the produced images in the context of painting style transfer. Indeed, transferring the style of a painting is a complex task involving features at different scales, from the color palette and compositional style to the fine brushstrokes and texture of the canvas. This paper provides a solution to solve the original global optimization for ultra-high resolution (UHR) images, enabling multiscale NST at unprecedented image sizes. This is achieved by spatially localizing the computation of each forward and backward passes through the VGG network. Extensive qualitative and quantitative comparisons, as well as a \textcolor{coverletter}{perceptual study}, show that our method produces style transfer of unmatched quality for such high-resolution painting styles. By a careful comparison, we show that state-of-the-art fast methods are still prone to artifacts, thus suggesting that fast painting style transfer remains an open problem. Source code is available at https://github.com/bgalerne/scaling_painting_style_transfer.

Hyam Omar Ali, Antoine Crosnier, Romain Abraham et al.

Sentinel-5P (S5P) plays a critical role in atmospheric monitoring; however, its spatial resolution limits fine-scale analysis. Existing super-resolution (SR) approaches rely on supervised learning with synthetic low-resolution (LR) data, since true high-resolution (HR) data do not exist, limiting their applicability to real observations. We propose a self-supervised hyperspectral SR framework for S5P that enables training without HR ground truth. The method combines Stein's Unbiased Risk Estimator (SURE) with an equivariant imaging constraint, incorporating the S5P degradation operator and noise statistics derived from signal-to-noise ratio (SNR) metadata. We also introduce depthwise separable convolution U-Net architectures designed for efficiency and spectral fidelity. The framework is evaluated in two settings: (i) LR-HR, where synthetic LR data are used for direct comparison with supervised learning, and (ii) GT-SHR, where super-resolved images surpass the native spatial resolution without HR reference. Results across multiple bands show that self-supervised models achieve performance comparable to supervised methods while maintaining strong consistency. Qualitative analysis shows improved spatial detail over bicubic interpolation, and validation with EMIT data confirms that reconstructed structures are physically meaningful. Code is available at https://github.com/hyamomar/Sentinel-5P-Super-Resolution/tree/main/self_supervised

Ngoc Long Nguyen, Jérémy Anger, Lara Raad et al.

In this work, we study the problem of single-image super-resolution (SISR) of Sentinel-2 imagery. We show that thanks to its unique sensor specification, namely the inter-band shift and alias, that deep-learning methods are able to recover fine details. By training a model using a simple $L_1$ loss, results are free of hallucinated details. For this study, we build a dataset of pairs of images Sentinel-2/PlanetScope to train and evaluate our super-resolution (SR) model.

Valéry Dewil, Zhe Zheng, Arnaud Barral et al.

MIMO (multiple input, multiple output) approaches are a recent trend in neural network architectures for video restoration problems, where each network evaluation produces multiple output frames. The video is split into non-overlapping stacks of frames that are processed independently, resulting in a very appealing trade-off between output quality and computational cost. In this work we focus on the low-latency setting by limiting the number of available future frames. We find that MIMO architectures suffer from problems that have received little attention so far, namely (1) the performance drops significantly due to the reduced temporal receptive field, particularly for frames at the borders of the stack, (2) there are strong temporal discontinuities at stack transitions which induce a step-wise motion artifact. We propose two simple solutions to alleviate these problems: recurrence across MIMO stacks to boost the output quality by implicitly increasing the temporal receptive field, and overlapping of the output stacks to smooth the temporal discontinuity at stack transitions. These modifications can be applied to any MIMO architecture. We test them on three state-of-the-art video denoising networks with different computational cost. The proposed contributions result in a new state-of-the-art for low-latency networks, both in terms of reconstruction error and temporal consistency. As an additional contribution, we introduce a new benchmark consisting of drone footage that highlights temporal consistency issues that are not apparent in the standard benchmarks.

Emile Pierret, Bruno Galerne

Diffusion or score-based models recently showed high performance in image generation. They rely on a forward and a backward stochastic differential equations (SDE). The sampling of a data distribution is achieved by numerically solving the backward SDE or its associated flow ODE. Studying the convergence of these models necessitates to control four different types of error: the initialization error, the truncation error, the discretization error and the score approximation. In this paper, we theoretically study the behavior of diffusion models and their numerical implementation when the data distribution is Gaussian. Our first contribution is to derive the analytical solutions of the backward SDE and the probability flow ODE and to prove that these solutions and their discretizations are all Gaussian processes. Our second contribution is to compute the exact Wasserstein errors between the target and the numerically sampled distributions for any numerical scheme. This allows us to monitor convergence directly in the data space, while experimental works limit their empirical analysis to Inception features. An implementation of our code is available online.

Bruno Galerne, Jianling Wang, Lara Raad et al.

Applying style transfer to a full 3D environment is a challenging task that has seen many developments since the advent of neural rendering. 3D Gaussian splatting (3DGS) has recently pushed further many limits of neural rendering in terms of training speed and reconstruction quality. This work introduces SGSST: Scaling Gaussian Splatting Style Transfer, an optimization-based method to apply style transfer to pretrained 3DGS scenes. We demonstrate that a new multiscale loss based on global neural statistics, that we name SOS for Simultaneously Optimized Scales, enables style transfer to ultra-high resolution 3D scenes. Not only SGSST pioneers 3D scene style transfer at such high image resolutions, it also produces superior visual quality as assessed by thorough qualitative, quantitative and perceptual comparisons.

Emile Pierret, Bruno Galerne

Used as priors for Bayesian inverse problems, diffusion models have recently attracted considerable attention in the literature. Their flexibility and high variance enable them to generate multiple solutions for a given task, such as inpainting, super-resolution, and deblurring. However, several unresolved questions remain about how well they perform. In this article, we investigate the accuracy of these models when applied to a Gaussian data distribution for deblurring. Within this constrained context, we are able to precisely analyze the discrepancy between the theoretical resolution of inverse problems and their resolution obtained using diffusion models by computing the exact Wasserstein distance between the distribution of the diffusion model sampler and the ideal distribution of solutions to the inverse problem. Our findings allow for the comparison of different algorithms from the literature.

Jorge Gutierrez, Julien Rabin, Bruno Galerne et al.

This paper describes a novel approach for on demand volumetric texture synthesis based on a deep learning framework that allows for the generation of high quality 3D data at interactive rates. Based on a few example images of textures, a generative network is trained to synthesize coherent portions of solid textures of arbitrary sizes that reproduce the visual characteristics of the examples along some directions. To cope with memory limitations and computation complexity that are inherent to both high resolution and 3D processing on the GPU, only 2D textures referred to as "slices" are generated during the training stage. These synthetic textures are compared to exemplar images via a perceptual loss function based on a pre-trained deep network. The proposed network is very light (less than 100k parameters), therefore it only requires sustainable training (i.e. few hours) and is capable of very fast generation (around a second for $256^3$ voxels) on a single GPU. Integrated with a spatially seeded PRNG the proposed generator network directly returns an RGB value given a set of 3D coordinates. The synthesized volumes have good visual results that are at least equivalent to the state-of-the-art patch based approaches. They are naturally seamlessly tileable and can be fully generated in parallel.

Valentin De Bortoli, Agnes Desolneux, Alain Durmus et al.

Recent years have seen the rise of convolutional neural network techniques in exemplar-based image synthesis. These methods often rely on the minimization of some variational formulation on the image space for which the minimizers are assumed to be the solutions of the synthesis problem. In this paper we investigate, both theoretically and experimentally, another framework to deal with this problem using an alternate sampling/minimization scheme. First, we use results from information geometry to assess that our method yields a probability measure which has maximum entropy under some constraints in expectation. Then, we turn to the analysis of our method and we show, using recent results from the Markov chain literature, that its error can be explicitly bounded with constants which depend polynomially in the dimension even in the non-convex setting. This includes the case where the constraints are defined via a differentiable neural network. Finally, we present an extensive experimental study of the model, including a comparison with state-of-the-art methods and an extension to style transfer.

Claire Launay, Bruno Galerne, Agnès Desolneux

Determinantal point processes (DPPs) enable the modeling of repulsion: they provide diverse sets of points. The repulsion is encoded in a kernel $K$ that can be seen as a matrix storing the similarity between points. The diversity comes from the fact that the inclusion probability of a subset is equal to the determinant of a submatrice of $K$. The exact algorithm to sample DPPs uses the spectral decomposition of $K$, a computation that becomes costly when dealing with a high number of points. Here, we present an alternative exact algorithm in the discrete setting that avoids the eigenvalues and the eigenvectors computation. Instead, it relies on Cholesky decompositions. This is a two steps strategy: first, it samples a Bernoulli point process with an appropriate distribution, then it samples the target DPP distribution through a thinning procedure. Not only is the method used here innovative, but this algorithm can be competitive with the original algorithm or even faster for some applications specified here.