Lightweight Optimal-Transport Harmonization on Edge Devices

Lightweight Optimal-Transport Harmonization on Edge Devices

Maria Larchenko, Dmitry Guskov, Alexander Lobashev, Georgy Derevyanko
AAAI 2026 2026
optimal-transport edge-computing mobile-ai

Color harmonization adjusts the colors of an inserted object so that it perceptually matches the surrounding image, resulting in a seamless composite. The harmonization problem naturally arises in augmented reality...

Training-Free Voice Conversion with Factorized Optimal Transport

Training-Free Voice Conversion with Factorized Optimal Transport

Alexander Lobashev, Assel Yermekova, Maria Larchenko
Interspeech 2025 2025
optimal-transport voice-conversion audio-processing

This paper introduces Factorized MKL-VC, a training-free modification for kNN-VC pipeline. In contrast with original pipeline, our algorithm performs high quality any-to-any cross-lingual voice conversion with only 5 second of...

Hessian Geometry of Latent Space in Generative Models

Hessian Geometry of Latent Space in Generative Models

Alexander Lobashev, Dmitry Guskov, Maria Larchenko, Mikhail Tamm
ICML 2025 2025
latent-space-geometry diffusion-models fisher-information

This paper presents a novel method for analyzing the latent space geometry of generative models, including statistical physics models and diffusion models, by reconstructing the Fisher information metric. The method...

Color Transfer with Modulated Flows

Color Transfer with Modulated Flows

Maria Larchenko, Alexander Lobashev, Dmitry Guskov, Vladimir Vladimirovich Palyulin
AAAI 2025 2025
color-transfer optimal-transport rectified-flows

In this work, we introduce Modulated Flows (ModFlows), a novel approach for color transfer between images based on rectified flows. The primary goal of color transfer is to adjust the...

Color Conditional Generation with Sliced Wasserstein Guidance

Color Conditional Generation with Sliced Wasserstein Guidance

Alexander Lobashev, Maria Larchenko, Dmitry Guskov
NeurIPS 2025 2025
color-transfer diffusion-models generative-models

We propose SW-Guidance, a training-free approach for image generation conditioned on the color distribution of a reference image. While it is possible to generate an image with fixed colors by...

Caffarelli Regularity and Hierarchical Phase Boundaries in Diffusion Model Latent Space

Caffarelli Regularity and Hierarchical Phase Boundaries in Diffusion Model Latent Space

Alexander Lobashev, Dmitry Guskov, Victor Borruat, Maria Larchenko
Under Review 2025
diffusion-models optimal-transport latent-space-geometry

This paper investigates the geometric structure of diffusion model latent spaces through the lens of optimal transport theory and Caffarelli regularity. We reveal hierarchical phase boundaries that organize the latent...

fBm-Based Generative Inpainting for the Reconstruction of Chromosomal Distances

fBm-Based Generative Inpainting for the Reconstruction of Chromosomal Distances

Alexander Lobashev, Dmitry Guskov, Kirill Polovnikov
NeurIPS 2024 2024
diffusion-models inpainting bioinformatics

Fractional Brownian motion (fBm) features both randomness and strong scale-free correlations, challenging generative models to reproduce the intrinsic memory characterizing the underlying stochastic process. Here we examine a zoo of...

Creative Any-to-Cloth Generation with Virtual Try-On Pipelines

Creative Any-to-Cloth Generation with Virtual Try-On Pipelines

Alexander Lobashev, Maria Larchenko, Dmitry Guskov
Multimodal AI TTIC 2024 2024
virtual-try-on fashion-ai vision-language-models

Virtual try-on (VITON) technology changes the way we shop by allowing customers to see how clothes look on them using augmented reality and artificial intelligence. This technology is important for...