Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks106 from public data
- Generalized Intersection Over Union: A Metric and a Loss for Bounding Box Regression6,169
This paper introduces a generalized version of IoU ( GIoU) as a loss into the state-of-the art object detection frameworks, and shows a consistent improvement on their performance using both the standard, IoU based, and new, GIo U based, performance measures on popular object detection benchmarks.
- GMFlow: Learning Optical Flow via Global Matching583
A GMFlow framework is proposed, which consists of three main components: a customized Transformer for feature enhancement, a correlation and softmax layer for global feature matching, and a self-attention layer for flow propagation, which outperforms 31-refinements RAFT on the challenging Sintel benchmark, while using only one refinement and running faster.
- 544
- 497
- Unifying Flow, Stereo and Depth Estimation433
It is demonstrated that cross-attention enables integration of knowledge from another image via cross-view interactions, which greatly improves the quality of the extracted features and enables cross-task transfer.
- 204
- Embryo selection through artificial intelligence versus embryologists: a systematic review140
AI consistently outperformed the clinical teams in all the studies focused on embryo morphology and clinical outcome prediction during embryo selection assessment and predicted the likelihood of successful clinical pregnancy with greater accuracy than clinical embryologists, signifying greater reliability when compared to human prediction.
- JRMOT: A Real-Time 3D Multi-Object Tracker and a New Large-Scale Dataset113
JRMOT is a novel 3D MOT system that integrates information from RGB images and 3D point clouds to achieve real-time, state-of-the-art tracking performance and serves as first 3D tracking solution for the authors' benchmark.
- MARLIN: Masked Autoencoder for facial video Representation LearnINg109
The proposed framework, named MARLIN, is a facial video masked autoencoder that learns highly robust and generic facial embeddings from abundantly available non-annotated web crawled facial videos that can transfer across a variety of facial analysis tasks such as Facial Attribute Recognition (FAR), Facial Expression Recognition, DeepFake Detection, and Lip Synchronization.
- Attend and Discriminate105
This work proposes learning to exploit the latent relationships between multi-channel sensor modalities and specific activities, and incorporates a classification loss criterion to encourage minimal intra-class representation differences whilst maximising inter-class differences to achieve more discriminative features.
- Probabilistic Tracklet Scoring and Inpainting for Multiple Object Tracking99
A probabilistic autoregressive motion model is introduced to score tracklet proposals by directly measuring their likelihood, achieved by training the model to learn the underlying distribution of natural tracklets.
- Socially and Contextually Aware Human Motion and Pose Forecasting96
A novel framework to tackle both tasks of human motion (or trajectory) and body skeleton pose forecasting in a unified end-to-end pipeline is proposed and achieves a superior performance compared to several baselines on two social datasets.
- Online UAV Path Planning for Joint Detection and Tracking of Multiple Radio-Tagged Objects95
This paper derives a partially observable Markov decision process with a random finite set track-before-detect (TBD) multi-object filter, which also maintains a safe distance between the UAV and the objects of interest using a void probability constraint, and demonstrates the capability of the proposed approach to handle multiple radio-tagged objects subject to birth, death, and motion modes.
- Joint Learning of Social Groups, Individuals Action and Sub-group Activities in Videos90
This paper proposes an end-to-end trainable framework for the social task, and sets the state-of-the-art results on two widely adopted benchmarks for the traditional group activity recognition task.
- DrVideo: Document Retrieval Based Long Video Understanding89
DrVideo is proposed, a document-retrieval-based system designed for long video understanding that significantly outperforms existing LLM-based state-of-the-art methods on EgoSchema benchmark, MovieChat-1K benchmark, and the long split of Video-MME benchmark.
- TRiPOD: Human Trajectory and Pose Dynamics Forecasting in the Wild77
A novel TRajectory and POse Dynamics (nicknamed TRiPOD) method based on graph attentional networks to model the human-human and human-object interactions both in the input space and the output space (decoded future output).
- Looking Beyond Two Frames: End-to-End Multi-Object Tracking Using Spatial and Temporal Transformers74
MO3TR is presented: a truly end-to-end Transformer-based online multi-object tracking (MOT) framework that learns to handle occlusions, track initiation and termination without the need for an explicit data association module or any heuristics.
- JRDB-Act: A Large-scale Dataset for Spatio-temporal Action, Social Group and Activity Detection70
This paper introduces JRDB-Act, an extension of the existing JRDB, which is captured by a social mobile manipulator and reflects a real distribution of human daily-life actions in a university campus environment and develops an end-to-end trainable pipeline to learn and infer these tasks, i.e. individual action and social group detection.
- Distributed Multi-Object Tracking Under Limited Field of View Sensors65
A novel distributed multi-target algorithm is developed that fuses local multiobject states instead of local multi-object densities and uses a novel label consensus approach that reduces label inconsistency, caused by movements of objects from one node’s limited FoV to another.
- 63
- ActiveRMAP: Radiance Field for Active Mapping And Planning53
This paper presents an RGB-only active vision framework using radiance field representation for active 3D reconstruction and planning in an online manner and suggests that the proposed method achieves competitive results compared to other offline methods and outperforms active reconstruction methods using NeRFs.
- JRDB-Pose: A Large-Scale Dataset for Multi-Person Pose Estimation and Tracking48
A thorough experimental study of state-of-the-art multi-person pose estimation and tracking methods on JRDB-Pose is conducted, showing that the dataset imposes new challenges for the existing methods.
- Deep Auto-Set46
This paper addresses HAR more naturally as a set prediction problem where the predictions are sets of ongoing activity elements with unfixed and unknown cardinality for the first time, and presents a novel HAR approach that learns to output activity sets using deep neural networks.
- Series2vec: similarity-based self-supervised representation learning for time series classification39
This work argues that time series analysis is fundamentally different in nature to either vision or natural language processing with respect to the forms of meaningful self-supervised learning tasks that can be defined, and introduces a novel approach called Series2Vec for self-supervised representation learning.
- Unsupervised Image Segmentation by Mutual Information Maximization and Adversarial Regularization39
This letter proposes a novel fully unsupervised semantic segmentation method, the so-called Information Maximization and Adversarial Regularization Segmentation (InMARS), Inspired by human perception which parses a scene into perceptual groups, rather than analyzing each pixel individually.
- DifFUSER: Diffusion Model for Robust Multi-Sensor Fusion in 3D Object Detection and BEV Segmentation35
DifFUSER, a novel approach that leverages diffusion models for multi-modal fusion in 3D object detection and BEV map segmentation, is introduced, able to refine or even synthesize sensor features in case of sensor malfunction, thereby improving the quality of the fused output.
- Accurate and Real-Time 3D Pedestrian Detection Using an Efficient Attentive Pillar Network35
This work introduces a stackable Pillar Aware Attention (PAA) module to enhance pillar feature extraction while suppressing noises in point clouds, and presents Mini-BiFPN, a small yet effective feature network that creates bidirectional information flow and multi-level cross-scale feature fusion to better integrate multi-resolution features.
- 34
- Protocon: Pseudo-Label Refinement via Online Clustering and Prototypical Consistency for Efficient Semi-Supervised Learning34
Protocon is a novel SSL method aimed at the less-explored label-scarce SSL where such methods usually underperform and delivers significant gains and faster convergence over state-of-the-art across 5 datasets, including CIFARs, ImageNet and DomainNet.
- ODAM: Object Detection, Association, and Mapping using Posed RGB Video34
ODAM, a system for 3D Object Detection, Association, and Mapping using posed RGB videos, relies on a deep learning front-end to detect 3D objects from a given RGB frame and associate them to a global object-based map using a graph neural network (GNN).
- How Trustworthy are Performance Evaluations for Basic Vision Tasks?33
A notion of trustworthiness for performance criteria is suggested, which requires robustness to parameters for reliability, contextual meaningfulness in sanity tests, and consistency with mathematical requirements such as the metric properties.
- Hier-SLAM: Scaling-Up Semantics in SLAM with a Hierarchically Categorical Gaussian Splatting32
Hier-SLAM is proposed, a semantic 3D Gaussian Splatting SLAM method featuring a novel hierarchical categorical representation that encodes semantic information in a compact form into 3D Gaussian Splatting, leveraging the capabilities of large language models (LLMs).
- SoMoFormer: Multi-Person Pose Forecasting with Transformers31
SoMoFormer outperforms state-of-the-art methods for long-term motion prediction on the SoMoF benchmark as well as the CMU-Mocap and MuPoTS-3D datasets.
- JRDB-Social: A Multifaceted Robotic Dataset for Understanding of Context and Dynamics of Human Interactions Within Social Groups30
This paper intro-duces JRDB-Social, an extension of JRDB designed to fill gaps in human understanding across diverse indoor and outdoor social contexts, and provides annotations at three levels: individual attributes, intra-group in-teractions, and social group context.
- 29
- Multi-Objective Multi-Agent Planning for Discovering and Tracking Multiple Mobile Objects27
It is proved that the proposed multi-objective value function is a monotone submodular set function, which admits low-cost suboptimal solutions via greedy search with a tight optimality bound.
- Learning of Global Objective for Network Flow in Multi-Object Tracking26
This paper proposes a novel differentiable framework that ties training and inference to-gether during learning by solving a bi-level optimization problem, where the lower-level solves a linear program and the upper-level contains a loss function that incorpo-rates global tracking result.
- MOLTR: Multiple Object Localization, Tracking and Reconstruction From Monocular RGB Videos25
MOLTR is presented, a solution to object-centric mapping using only monocular image sequences and camera poses that is able to localize, track and reconstruct multiple rigid objects in an online fashion when a RGB camera captures a video of the surrounding.
- Knowledge Combination to Learn Rotated Detection without Rotated Annotation22
A framework that allows the model to predict precise rotated boxes only requiring cheaper axis-aligned annotation of the target dataset only and is able to solve the more detailed task in the target domain, without additional computation overhead during inference is proposed.
- Learn to Predict Sets Using Feed-Forward Neural Networks22
This paper presents a novel approach for learning to predict sets with unknown permutation and cardinality using deep neural networks and demonstrates the validity of the set formulations on relevant vision problems such as multi-label image classification and a complex CAPTCHA test.
- AerOSeg: Harnessing SAM for Open-Vocabulary Segmentation in Remote Sensing Images21
AerOSeg, a novel OVS approach for remote sensing data, inspired by the success of the Segment Anything Model (SAM) in diverse domains, outperforms state-of-the-art open-vocabulary segmentation methods and validated on three benchmark remote sensing datasets.
- Deep learning classification integrating embryo images with associated clinical information from ART cycles18
This study demonstrates that AI for IVF application can increase prediction performance by integrating blastocyst images with patient clinical information, and finds that female and male age to be the most clinical factors, whilst Trophectoderm to be the most important blastocyst feature.
- NEUSIS: A Compositional Neuro-Symbolic Framework for Autonomous Perception, Reasoning, and Planning in Complex UAV Search Missions18
Experimental results from simulated urban search missions using AirSim and Unreal Engine show that NEUSIS outperforms state-of-the-art baselines for both perception and planning, making it a promising solution for autonomous UAV systems in search missions.
- 17
- How Well Can Vision Language Models See Image Details?16
This research reveals that incorporating pixel value prediction as one of the VLM pre-training tasks and vision encoder adaptation markedly boosts VLM performance on downstream image-language understanding tasks requiring detailed image perception and video game decision making.
- 15
- Naver: a Neuro-Symbolic Compositional Automaton for Visual Grounding with Explicit Logic Reasoning14
NAVER is a compositional visual grounding method that integrates explicit probabilistic logic reasoning within a finite-state automaton, equipped with a self-correcting mechanism, that improves ro-bustness and interpretability in inference through explicit logic reasoning.
- DWIM: Towards Tool-Aware Visual Reasoning via Discrepancy-Aware Workflow Generation & Instruct-Masking Tuning12
DWIM is proposed, which guides the model to only clone effective actions, enabling the generation of more practical solutions in VR, and achieves state-of-the-art performance across various VR tasks, exhibiting strong generalization on multiple widely-used datasets.
- 12
- 10
- Real-Time Trajectory-Based Social Group Detection10
This approach explores the impact of motion trajectory on social grouping and utilizes a novel, reliable, and fast data-driven method that is significantly faster, with up to 12x faster inference times compared to state-of-the-art methods under the same computation resources.
- Meta Learning with Differentiable Closed-form Solver for Fast Video Object Segmentation10
This paper tackles the task of fast segmenting all pixels of a target object in all frames, given the annotation mask in the first frame, as a meta-learning problem, where the base learner grasping the semantic scene understanding for a general type of objects, and the meta learner quickly adapting the appearance of the target object with a few examples.
- How Trustworthy are the Existing Performance Evaluations for Basic Vision Tasks10
This paper examines performance evaluation criteria for basic vision tasks namely, object detection, instance-level segmentation and multi-object tracking, and advocates the use of criteria that are consistent with mathematical requirements, contextually meaningful in sanity tests, and robust to hyper-parameters for reliability.
- Acoustic-to-hyper-spectral: real-time perimeter intrusion detection system monitoring through learnable filters and hyper-spectral image generation from distributed acoustic sensing systems9
What the authors believe to be three novel signal processing methods are proposed to train filters for automatically selecting frequency bands from the power spectrum and generating hyper-spectral images from the data gathered by the DAS system without expert knowledge.
- Mobile-VideoGPT: Fast and Accurate Model for Mobile Video Understanding9
The proposed Mobile-VideoGPT, an efficient multimodal framework designed to operate with fewer than a billion parameters, consists of lightweight dual visual encoders, efficient projectors, and a small language model (SLM), enabling real-time through-put.
- 8
- JRDB-Traj: A Dataset and Benchmark for Trajectory Forecasting in Crowds7
A novel dataset for end-to-end trajectory forecasting, facilitating the evaluation of models in scenarios involving less-than-ideal preceding modules such as tracking, and introduces a novel metric for assessing trajectory forecasting models in real-world scenarios where ground-truth identities are inaccessible.
- Marginalized Generalized IoU (MGIoU): A Unified Objective Function for Optimizing Any Convex Parametric Shapes5
Marginalized Generalized IoU is introduced, a novel loss function that overcomes challenges by projecting structured convex shapes onto their unique shape Normals to compute one-dimensional normalized GIoU and extends MGIoU to MGIoU+ that supports optimizing unstructured convex shapes.
- Social-MAE: Social Masked Autoencoder for Multi-Person Motion Representation Learning5
Social-MAE, a simple yet effective transformer-based masked autoencoder framework for multi-person human motion data, uses masked modeling to pre-train the encoder to reconstruct masked human joint trajectories, enabling it to learn generalizable representations of motion in human crowded scenes.
- 5
- Physically Plausible 3D Human-Scene Reconstruction From Monocular RGB Image Using an Adversarial Learning Approach5
This letter proposes using an implicit feature representation of the scene elements to distinguish a physically plausible alignment of humans and objects from an implausible one and adversarially trained model to produce a per-frame 3D reconstruction of thescene that abides by the physical laws and constraints.
- 5
- 5
- 5
- An Empirical Study on How Video-LLMs Answer Video Questions4
This is the first work to systematically uncover how Video-LLMs internally process and understand video content, offering interpretability and efficiency perspectives for future research.
- Physics-Grounded Motion Forecasting via Equation Discovery for Trajectory-Guided Image-to-Video Generation4
This work introduces a novel framework that integrates symbolic regression (SR) and trajectory-guided image-to-video (I2V) models for physics-grounded video forecasting and recovers ground-truth analytical equations and improves the physical alignment of generated videos over baseline methods.
- ECORE: Energy-Conscious Optimized Routing for Deep Learning Models at the Edge3
This work proposes ECORE, a framework that integrates multiple dynamic routing strategies, including a novel estimation-based techniques and an innovative greedy selection algorithm, to direct image processing requests to the most suitable edge device–model pair and demonstrates that this framework can reduce energy consumption and latency by 35% and 49%, respectively.
- MATA: A Trainable Hierarchical Automaton System for Multi-Agent Visual Reasoning3
MATA (Multi-Agent hierarchical Trainable Automaton), a multi-agent system presented as a hierarchical finite-state automaton for visual reasoning whose top-level transitions are chosen by a trainable hyper agent, achieves the state of theart results compared with monolithic and compositional baselines.
- 3
- 2
- AerOSeg++: Scale-Aware and Texture-Guided Open-Vocabulary Segmentation with SAM Features for Remote Sensing Images2
AerOSeg++ is introduced, an open-vocabulary segmentation method in remote sensing, focusing on scale-invariant feature learning, that has shown significant performance gains compared to the state-of-the-art open-vocabulary methods when evaluated on three benchmark datasets for remote sensing.
- CalibAnyView: Beyond Single-View Camera Calibration in the Wild2
CalibAnyView is a framework that unifies single- and sparse multi-view calibration by enforcing that consistency inside the network: a transformer with cross-view attention predicts camera-model-agnostic perspective fields, followed by a multi-view optimization that fuses them into shared intrinsics and per-view gravity directions.
- Sample-Efficient Learning from Agent Experience2
Compared with classical reinforcement-learning baselines, in-context learning from trial-and-error experience followed by Experience Distillation matches their performance with at least \(9.6\times\) fewer environment samples.
- 2
- LAVA:Label-efficient Visual Learning and Adaptation2
LAVA learns self-supervised visual representations on the source dataset and ground them using class label semantics to overcome transfer collapse problems associated with supervised pretraining and maximises the gains from unlabelled target data via a novel method which uses multi-crop augmentations to obtain highly robust pseudo-labels.
- 2
- 2
- A Multi-Modal Neuro-Symbolic Approach for Spatial Reasoning-Based Visual Grounding in Robotics1
A novel neuro_symbolic framework that integrates both panoramic-image and 3D point cloud information, combining neural perception with symbolic reasoning to explicitly model spatial and logical relationships is proposed.
- Improving Visual Perception of a Social Robot for Controlled and In-the-wild Human-robot Interaction1
- 1
- 1
- dinov3.seg: Open-Vocabulary Semantic Segmentation with DINOv3–
This work designs a task-specific architecture tailored to this backbone of OVSS, systematically adapting established design principles from prior open-vocabulary segmentation work, and proposes a high-resolution local-global inference strategy based on sliding-window aggregation, which preserves spatial detail while maintaining global context.
- JRDB-Pose3D: A Multi-person 3D Human Pose and Shape Estimation Dataset for Robotics–
JRDB-Pose3D inherits all available annotations from the JRDB dataset, such as 2D pose, information about social grouping, activities, and interactions, full-scene semantic masks with consistent human- and object-level tracking, and detailed annotations for each individual, making it a holistic dataset for a wide range of downstream perception and human-centric understanding tasks.
- –
- –
- Mini-BEHAVIOR-Gran: Revealing U-Shaped Effects of Instruction Granularity on Language-Guided Embodied Agents–
Mini-BEHAVIOR-Gran is introduced, a new benchmark for controlled studies of instruction granularity that extends Mini-BEHAVIOR with multiple instruction variants per task, ranging from high-level goal descriptions to step-by-step guidance, and reveals a non-monotonic U-shaped relationship between instruction granularity and performance.
- –
- –
- –
- ARIS: Agentic and Relationship Intelligence System for Social Robots–
ARIS (Agentic and Relationship Intelligence System), an agentic AI framework that unifies multimodal reasoning, a graph-based Social World Model, and retrieval-augmented generation (RAG) within a single modular architecture for social robots is presented.
- –
- Neurosymbolic Active Goal Recognition in Partially Observable Environments–
A neurosymbolic framework is introduced that extends the modeling capacity to account for heterogeneous actor behaviors and integrates neural models into the active goal recognition framework in two complementary ways, highlighting the promise of neurosymbolic methods for robust active goal recognition in complex, uncertain environments.
- –
- A Simple State Space Model Excels at Multivariate Time Series Classification–
This work presents the first systematic study spanning diagonal SSMs (S4D) and input-dependent SSMs (Mamba family) on large-scale TSC benchmarks, asking whether such complexity is necessary for top performance, and reveals a surprising finding: S4D consistently outperforms Mamba-based variants in both accuracy and efficiency.
- ASAP-Textured Gaussians: Enhancing Textured Gaussians with Adaptive Sampling and Anisotropic Parameterization–
The proposed ASAP Textured Gaussians, short for Adaptive Sampling and Anisotropic Parameterization, significantly improve the quality efficiency tradeoff, achieving high-fidelity rendering with far fewer texture parameters.
- –
- Asset-Driven Sematic Reconstruction of Dynamic Scene with Multi-Human-Object Interactions–
This work proposes a hybrid approach that effectively combines the advantages of 3D generative models for generating high-fidelity meshes of the scene elements, and helps maintain the object structures even under severe occlusion and can produce multiview and temporally consistent geometry.
- –
- –
- –
- –
- –
- –
- Predicting Topological Maps for Visual Navigation in Unexplored Environments–
This work describes a navigation system that uses the layout predictions to satisfy high-level goals more rapidly and accurately than the prior art and test the framework in Matterport3D to show more success and efficient navigation in unseen environments.
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.