Niculae Sebe

Full Professor

About

Contact & Online Profiles

The main research interests relate to the investigation and implementation of new techniques in the fields of computer vision and multimedia. Specifically, in computer vision, I address a large spectrum of themes including human-behavior analysis, action recognition, 2D/3D object detection, large-scale event detection and video analysis, etc. In multimedia, my research focuses on three aspects: multimedia information retrieval, social signals processing and affective computing.

The specific research topics include cross media retrieval, multi-modal learning, social media analysis, emotion recognition, multimodal brain-computer interfaces, etc.

Papers

CVPR 2025

About

Contact & Online Profiles

Papers

Multi-focal Conditioned Latent Diffusion for Person Image Synthesis

Gyrogroup Batch Normalization

Understanding Matrix Function Normalizations in Covariance Pooling from the Lens of Riemannian Geometry

UVMap-ID: A Controllable and Personalized UV Map Generative Model

OpenBias: Open-set Bias Detection in Text-to-Image Generative Models

PAIR-Diffusion: A Comprehensive Multimodal Object-Level Image Editor

Riemannian Multinomial Logistics Regression for SPD Neural Networks

A Lie Group Approach to Riemannian Batch Normalization

Democratizing Fine-grained Visual Recognition with Large Language Models

Adaptive Log-Euclidean Metrics for SPD Matrix Learning

Improving Fairness using Vision-Language Driven Image Augmentation

SpectralCLIP: Preventing Artifacts in Text-Guided Style Transfer from a Spectral Perspective

FedVQA: Personalized Federated Visual Question Answering over Heterogeneous Scenes

Multi-Domain Lifelong Visual Question Answering via Self-Critical Distillation

Attribute-preserving Face Dataset Anonymization via Latent Code Optimization

Dynamic Conceptional Contrastive Learning for Generalized Category Discovery

Dynamically Instance-Guided Adaptation: A Backward-Free Approach for Test-Time Domain Adaptive Semantic Segmentation

Masked Jigsaw Puzzle: A Versatile Position Embedding for Vision Transformers

PI-Trans: Parallel-ConvMLP and Implicit-Transformation Based GAN for Cross-View Image Translation

Householder Projector for Unsupervised Latent Semantics Discovery

StylerDALLE: Language-Guided Style Transfer Using a Vector-Quantized Tokenizer of a Large-Scale Generative Model

Latent Traversals in Generative Models as Potential Flows

Unleashing the Transferability Power of Unsupervised Pre-Training for Emotion Recognition in Masked and Unmasked Facial Images

A Memorizing and Generalizing Framework for Lifelong Person Re-Identification

Compositional Semantic Mix for Domain Adaptation in Point Cloud Segmentation

Towards Robust Person Re-Identification by Defending Against Universal Attackers

Flow Factorized Representation Learning

Overlap-guided Gaussian Mixture Model for Point Cloud Registration

Data Augmentation-free Unsupervised Learning for 3D Point Cloud Understanding

Low-budget label query through domain alignment enforcement

Hyperbolic Vision Transformers: Combining Improvements in Metric Learning

Novel Class Discovery in Semantic Segmentation

Predict, Prevent, and Evaluate: Disentangled Text-Driven Image Manipulation Empowered by Pre-Trained Vision-Language Model

3D-Aware Semantic-Guided Generative Model for Human Synthesis

Batch-efficient Eigen Decomposition for Small and Medium Matrices

Class-incremental Novel Class Discovery

CoSMix: Compositional Semantic Mix for Domain Adaptation in 3D LiDAR Segmentation

GIPSO: Geometrically Informed Propagation for Online Adaptation in 3D LiDAR Segmentation

Style-Hallucinated Dual Consistency Learning for Domain Generalized Semantic Segmentation.

Fast Differentiable Matrix Square Root

Temporal Alignment for History Representation in Reinforcement Learning

Unsupervised Domain Adaptation for Video Transformers in Action Recognition

Fast Differentiable Matrix Square Root and Inverse Square Root

On the Eigenvalues of Global Covariance Pooling for Fine-grained Visual Recognition

Orthogonal SVD Covariance Conditioning and Latent Disentanglement.

Joint Representation Learning and Keypoint Detection for Cross-view Geo-localization

ISF-GAN: An Implicit Style Function for High Resolution Image-to-Image Translation

Curriculum Learning: A Survey

Adversarial Style Augmentation for Domain Generalized Urban-Scene Segmentation

RankFeat: Rank-1 Feature Removal for Out-of-distribution Detection

Source-Free Open Compound Domain Adaptation in Semantic Segmentation

Dual-Head Contrastive Domain Adaptation for Video Action Recognition

Learning to Attack Real-World Models for Person Re-identification via Virtual-Guided Meta-Learning

Curriculum self-paced learning for cross-domain object detection

Coarse-to-Fine Gaze Redirection with Numerical and Pictorial Guidance

SF-UDA<sup>3D</sup>: Source-Free Unsupervised Domain Adaptation for LiDAR-Based 3D Object Detection

Dual In-painting Model for Unsupervised Gaze Correction and Animation in the Wild

Class-Aware Modality Mix and Center-Guided Metric Learning for Visible-Thermal Person Re-Identification

Learning How to Smile: Expression Video Generation with Conditional Adversarial Recurrent Nets

Latent World Models For Intrinsically Motivated Exploration

Whitening for Self-Supervised Representation Learning

Low-Budget Unsupervised Label Query through Domain Alignment Enforcement

Unsupervised domain adaptation using feature-whitening and consensus loss

Regularized Evolutionary Algorithm for Dynamic Neural Topology Search

Self-Paced Deep Learning for Weakly Supervised Object Detection

Every Smile is Unique: Landmark-Guided Diverse Smile Generation

Flexible Manifold Learning With Optimal Graph for Image and Video Representation

Recurrent Convolutional Shape Regression

Recurrent Face Aging with Hierarchical Autoregressive Memory