1 code implementation • 19 Mar 2025 • Yuanzhi Zhu, Xi Wang, Stéphane Lathuilière, Vicky Kalogeiton
Masked Diffusion Models (MDMs) have emerged as a powerful generative modeling technique.
1 code implementation • 12 Mar 2025 • Thomas De Min, Subhankar Roy, Stéphane Lathuilière, Elisa Ricci, Massimiliano Mancini
MIU minimizes the mutual information between model features and group information, achieving unlearning while reducing performance degradation in the dominant group of the forget set.
1 code implementation • 18 Aug 2024 • Muhammad Rameez Ur Rahman, Jhony H. Giraldo, Indro Spinelli, Stéphane Lathuilière, Fabio Galasso
In this paper, we introduce OVOSE, the first Open-Vocabulary Semantic Segmentation algorithm for Event cameras.
no code implementations • 29 Jul 2024 • Ekaterina Iakovleva, Fabio Pizzati, Philip Torr, Stéphane Lathuilière
We use a large language model (LLM) to decompose the input instruction into specific instructions, i. e. well-defined interventions to apply to the input image to satisfy the user's request.
1 code implementation • 17 Jul 2024 • Hamza Rami, Jhony H. Giraldo, Nicolas Winckler, Stéphane Lathuilière
Re-Identification systems (Re-ID) are crucial for public safety but face the challenge of having to adapt to environments that differ from their training distribution.
1 code implementation • 16 Jul 2024 • Thomas De Min, Subhankar Roy, Massimiliano Mancini, Stéphane Lathuilière, Elisa Ricci
To this extent, existing MU approaches assume complete or partial access to the training data, which can be limited over time due to privacy regulations.
1 code implementation • 24 May 2024 • Thomas De Min, Massimiliano Mancini, Stéphane Lathuilière, Subhankar Roy, Elisa Ricci
Since independent pathways in truly incremental scenarios will result in an explosion of computation due to the quadratically complex multi-head self-attention (MSA) operation in prompt tuning, we propose to reduce the original patch token embeddings into summarized tokens.
1 code implementation • 23 Feb 2024 • Hamza Rami, Jhony H. Giraldo, Nicolas Winckler, Stéphane Lathuilière
Our framework is based on the extraction of a support set composed of source images that maximizes the similarity with the target data.
Online unsupervised domain adaptation
Person Re-Identification
1 code implementation • CVPR 2024 • Yasser Benigmim, Subhankar Roy, Slim Essid, Vicky Kalogeiton, Stéphane Lathuilière
Domain Generalized Semantic Segmentation (DGSS) deals with training a model on a labeled source domain with the aim of generalizing to unseen domains during inference.
1 code implementation • 14 Dec 2023 • Imad Eddine Marouf, Subhankar Roy, Enzo Tartaglione, Stéphane Lathuilière
In this work, we study the problem of continual learning (CL) where the goal is to learn a model on a sequence of tasks, such that the data from the previous tasks becomes unavailable while learning on the current task data.
1 code implementation • 7 Nov 2023 • Imad Eddine Marouf, Enzo Tartaglione, Stéphane Lathuilière
Vision Transformers (ViTs) have become one of the dominant architectures in computer vision, and pre-trained ViT models are commonly adapted to new tasks via fine-tuning.
1 code implementation • 17 Oct 2023 • Imad Eddine Marouf, Subhankar Roy, Enzo Tartaglione, Stéphane Lathuilière
However, repeated fine-tuning on each task destroys the rich representations of the PTMs and further leads to forgetting previous tasks.
1 code implementation • 16 Oct 2023 • Marlène Careil, Matthew J. Muckley, Jakob Verbeek, Stéphane Lathuilière
We find that our model leads to reconstructions with state-of-the-art visual quality as measured by FID and KID.
1 code implementation • 20 Sep 2023 • Xiangyi Chen, Stéphane Lathuilière
We propose FADING, a novel approach to address Face Aging via DIffusion-based editiNG.
1 code implementation • ICCV 2023 • Giacomo Zara, Alessandro Conti, Subhankar Roy, Stéphane Lathuilière, Paolo Rota, Elisa Ricci
Source-Free Video Unsupervised Domain Adaptation (SFVUDA) task consists in adapting an action recognition model, trained on a labelled source dataset, to an unlabelled target dataset, without accessing the actual source data.
1 code implementation • 9 Jul 2023 • Goluck Konuko, Stéphane Lathuilière, Giuseppe Valenzise
We address the problem of efficiently compressing video for conferencing-type applications.
no code implementations • ICCV 2023 • Guillaume Couairon, Marlène Careil, Matthieu Cord, Stéphane Lathuilière, Jakob Verbeek
Large-scale text-to-image diffusion models have significantly improved the state of the art in generative image modelling and allow for an intuitive and powerful user interface to drive the image generation process.
no code implementations • 7 Apr 2023 • Marcelin Tworski, Stéphane Lathuilière
In this paper, we address the well-known image quality assessment problem but in contrast from existing approaches that predict image quality independently for every images, we propose to jointly model different images depicting the same content to improve the precision of quality estimation.
no code implementations • CVPR 2023 • Marlène Careil, Jakob Verbeek, Stéphane Lathuilière
The class affinity matrix is introduced as a first layer to the source model to make it compatible with the target label maps, and the source model is then further finetuned for the target domain.
1 code implementation • 31 Mar 2023 • Yasser Benigmim, Subhankar Roy, Slim Essid, Vicky Kalogeiton, Stéphane Lathuilière
Departing from the common notion of transferring only the target ``texture'' information, we leverage text-to-image diffusion models (e. g., Stable Diffusion) to generate a synthetic target dataset with photo-realistic images that not only faithfully depict the style of the target domain, but are also characterized by novel scenes in diverse contexts.
Data Augmentation
One-shot Unsupervised Domain Adaptation
+2
no code implementations • 23 Mar 2023 • Willi Menapace, Aliaksandr Siarohin, Stéphane Lathuilière, Panos Achlioptas, Vladislav Golyanik, Sergey Tulyakov, Elisa Ricci
Most captivatingly, our PGM unlocks the director's mode, where the game is played by specifying goals for the agents in the form of a prompt.
no code implementations • 25 Nov 2022 • Marlène Careil, Stéphane Lathuilière, Camille Couprie, Jakob Verbeek
To allow for more control, image synthesis can be conditioned on semantic segmentation maps that instruct the generator the position of objects in the image.
1 code implementation • 2 Nov 2022 • Louis Airale, Xavier Alameda-Pineda, Stéphane Lathuilière, Dominique Vaufreydaz
In this work, we address the task of unconditional head motion generation to animate still human faces in a low-dimensional semantic space from a single reference pose.
1 code implementation • 4 Oct 2022 • Yangsong Zhang, Subhankar Roy, Hongtao Lu, Elisa Ricci, Stéphane Lathuilière
In this work we address multi-target domain adaptation (MTDA) in semantic segmentation, which consists in adapting a single model from an annotated source dataset to multiple unannotated target datasets that differ in their underlying data distributions.
1 code implementation • 22 Jul 2022 • Guillermo Gomez-Trenado, Stéphane Lathuilière, Pablo Mesejo, Óscar Cordón
In this work, we propose a novel architecture for face age editing that can produce structural modifications while maintaining relevant details present in the original image.
2 code implementations • 20 Jul 2022 • Cristiano Saltori, Evgeny Krivosheev, Stéphane Lathuilière, Nicu Sebe, Fabio Galasso, Giuseppe Fiameni, Elisa Ricci, Fabio Poiesi
Our experiments show the effectiveness of our segmentation approach on thousands of real-world point clouds.
no code implementations • 9 May 2022 • Hamza Rami, Matthieu Ospici, Stéphane Lathuilière
Therefore, we present a new yet practical online setting for Unsupervised Domain Adaptation for person Re-ID with two main constraints: Online Adaptation and Privacy Protection.
Online unsupervised domain adaptation
Person Re-Identification
1 code implementation • CVPR 2022 • Willi Menapace, Stéphane Lathuilière, Aliaksandr Siarohin, Christian Theobalt, Sergey Tulyakov, Vladislav Golyanik, Elisa Ricci
We present Playable Environments - a new representation for interactive video generation and manipulation in space and time.
1 code implementation • ICCV 2021 • Enrico Fini, Enver Sangineto, Stéphane Lathuilière, Zhun Zhong, Moin Nabi, Elisa Ricci
In this paper, we study the problem of Novel Class Discovery (NCD).
Ranked #3 on
Novel Object Detection
on LVIS v1.0 val
1 code implementation • ICCV 2021 • Pierfrancesco Ardino, Marco De Nadai, Bruno Lepri, Elisa Ricci, Stéphane Lathuilière
This paper introduces Click to Move (C2M), a novel framework for video generation where the user can control the motion of the synthesized video through mouse clicks specifying simple object trajectories of the key objects in the scene.
no code implementations • 12 Jul 2021 • Enzo Tartaglione, Stéphane Lathuilière, Attilio Fiandrotti, Marco Cagnazzo, Marco Grangetto
We formulate the entropy of a quantized artificial neural network as a differentiable function that can be plugged as a regularization term into the cost function minimized by gradient descent.
1 code implementation • CVPR 2021 • Willi Menapace, Stéphane Lathuilière, Sergey Tulyakov, Aliaksandr Siarohin, Elisa Ricci
This paper introduces the unsupervised learning problem of playable video generation (PVG).
no code implementations • 11 Jan 2021 • The-Phuc Nguyen, Stéphane Lathuilière, Elisa Ricci
Therefore, we propose to increase the network capacity by using an adaptive graph structure.
1 code implementation • 1 Dec 2020 • Goluck Konuko, Giuseppe Valenzise, Stéphane Lathuilière
In this work we propose a novel deep learning approach for ultra-low bitrate video compression for video conferencing applications.
no code implementations • 23 Sep 2020 • Sylvain Guy, Stéphane Lathuilière, Pablo Mesejo, Radu Horaud
Visual voice activity detection (V-VAD) uses visual features to predict whether a person is speaking or not.
no code implementations • 21 Sep 2020 • Marcelin Tworski, Stéphane Lathuilière, Salim Belkarfa, Attilio Fiandrotti, Marco Cagnazzo
In this work, we tackle the problem of estimating a camera capability to preserve fine texture details at a given lighting condition.
1 code implementation • ECCV 2020 • Willi Menapace, Stéphane Lathuilière, Elisa Ricci
While unsupervised domain adaptation methods based on deep architectures have achieved remarkable success in many computer vision tasks, they rely on a strong assumption, i. e. labeled source data must be available.
1 code implementation • ECCV 2020 • Enrico Fini, Stéphane Lathuilière, Enver Sangineto, Moin Nabi, Elisa Ricci
Continual Learning (CL) aims to develop agents emulating the human ability to sequentially learn new tasks while being able to retain knowledge obtained from past experiences.
2 code implementations • 7 Apr 2020 • Aliaksandr Siarohin, Subhankar Roy, Stéphane Lathuilière, Sergey Tulyakov, Elisa Ricci, Nicu Sebe
To overcome this limitation, we propose a self-supervised deep learning method for co-part segmentation.
Ranked #3 on
Unsupervised Human Pose Estimation
on Tai-Chi-HD
3 code implementations • NeurIPS 2019 • Aliaksandr Siarohin, Stéphane Lathuilière, Sergey Tulyakov, Elisa Ricci, Nicu Sebe
To achieve this, we decouple appearance and motion information using a self-supervised formulation.
Ranked #1 on
Video Reconstruction
on Tai-Chi-HD
1 code implementation • 17 Sep 2019 • Andrea Pilzer, Stéphane Lathuilière, Dan Xu, Mihai Marian Puscas, Elisa Ricci, Nicu Sebe
Extensive experiments on the publicly available datasets KITTI, Cityscapes and ApolloScape demonstrate the effectiveness of the proposed model which is competitive with other unsupervised deep learning methods for depth prediction.
no code implementations • ICCV 2019 • Rodrigo Berriel, Stéphane Lathuilière, Moin Nabi, Tassilo Klein, Thiago Oliveira-Santos, Nicu Sebe, Elisa Ricci
To implement this idea we derive specialized deep models for each domain by adapting a pre-trained architecture but, differently from other methods, we propose a novel strategy to automatically adjust the computational complexity of the network.
no code implementations • 7 May 2019 • Stéphane Lathuilière, Enver Sangineto, Aliaksandr Siarohin, Nicu Sebe
We present a generalization of the person-image generation task, in which a human image is generated conditioned on a target pose and a set X of source appearance images.
1 code implementation • 30 Apr 2019 • Aliaksandr Siarohin, Stéphane Lathuilière, Enver Sangineto, Nicu Sebe
Specifically, given an image xa of a person and a target pose P(xb), extracted from a different image xb, we synthesize a new image of that person in pose P(xb), while preserving the visual details in xa.
no code implementations • 17 Apr 2019 • Zhen-Yu Zhang, Stéphane Lathuilière, Andrea Pilzer, Nicu Sebe, Elisa Ricci, Jian Yang
Our proposal is evaluated on the wellestablished KITTI dataset, where we show that our online method is competitive withstate of the art algorithms trained in a batch setting.
no code implementations • CVPR 2019 • Andrea Pilzer, Stéphane Lathuilière, Nicu Sebe, Elisa Ricci
Therefore, recent works have proposed deep architectures for addressing the monocular depth prediction task as a reconstruction problem, thus avoiding the need of collecting ground-truth depth.
no code implementations • 28 Feb 2019 • Benoit Massé, Stéphane Lathuilière, Pablo Mesejo, Radu Horaud
In this paper we address the problems of detecting objects of interest in a video and of estimating their locations, solely from the gaze directions of people present in the video.
1 code implementation • CVPR 2019 • Aliaksandr Siarohin, Stéphane Lathuilière, Sergey Tulyakov, Elisa Ricci, Nicu Sebe
This is achieved through a deep architecture that decouples appearance and motion information.
no code implementations • ECCV 2018 • Stéphane Lathuilière, Pablo Mesejo, Xavier Alameda-Pineda, Radu Horaud
In this paper, we address the problem of how to robustly train a ConvNet for regression, or deep robust regression.
2 code implementations • 22 Mar 2018 • Stéphane Lathuilière, Pablo Mesejo, Xavier Alameda-Pineda, Radu Horaud
Deep learning revolutionized data science, and recently its popularity has grown exponentially, as did the amount of papers employing deep networks.
no code implementations • 18 Nov 2017 • Stéphane Lathuilière, Benoit Massé, Pablo Mesejo, Radu Horaud
Our approach enables a robot to learn and to adapt its gaze control strategy for human-robot interaction neither with the use of external sensors nor with human supervision.