Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileAcademic lineage
Possible advisorsa guess from early papers, not confirmed
- Luc Van GoolSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works109 from public data
- Weakly Supervised Deep Detection Networks851
This paper proposes a weakly supervised deep detection architecture that modifies one such network to operate at the level of image regions, performing simultaneously region selection and classification.
- 604
- 558
- Dataset Condensation with Distribution Matching507
This work proposes a simple yet effective method that synthesizes condensed images by matching feature distributions of the synthetic and original training images in many sampled embedding spaces and significantly reduces the synthesis cost while achieving comparable or better performance.
- 452
- Efficient Parametrization of Multi-domain Deep Neural Networks391
This paper proposes to consider universal parametric families of neural networks, which still contain specialized problem-specific models, but differing only by a small number of parameters, and shows that these universal parametrization are very effective for transfer learning, where they outperform traditional fine-tuning techniques.
- 365
- 261
- 231
- RenderDiffusion: Image Diffusion for 3D Reconstruction, Inpainting and Generation217
RenderDiffusion is presented, the first diffusion model for 3D generation and inference, trained using only monocular 2D supervision, that generates and renders an intermediate three-dimensional representation of a scene in each denoising step and allows for 2D inpainting to edit 3D scenes.
- Unsupervised Learning of Object Landmarks by Factorized Spatial Embeddings173
This paper proposes a novel unsupervised approach that can discover and learn landmarks in object categories, thus characterizing their structure, and shows that the learned landmarks establish meaningful correspondences between different object instances in a category without having to impose this requirement explicitly.
- Cross-domain Few-shot Learning with Task-specific Adapters170
This work proposes to learn task- specific weights from scratch directly on a small support set, in contrast to dynamically estimating them, and shows that task-specific weights through parametric adapters in matrix form with residual connections to multiple intermediate layers of a backbone network significantly improves the per-formance of the state-of-the-art models in the Meta-Dataset benchmark with minor additional cost.
- CAFE: Learning to Condense Dataset by Aligning Features165
This paper proposes a novel scheme to Condense dataset by Aligning FEatures (CAFE), which explicitly attempts to preserve the real-feature distribution as well as the discriminant power of the resulting synthetic set, lending itself to strong generalization capability to various architectures.
- Weakly Supervised Object Detection with Posterior Regularization133
A method based on features extracted from a Convolutional Neural Network and latent SVM that can represent and exploit the presence of multiple object instances in an image that outperforms the state-of-the-art in weakly-supervised object detection and object classification on the Pascal VOC 2007 dataset.
- Universal Representation Learning from Multiple Domains for Few-shot Classification117
URL is presented, which learns a single set of universal visual representations by distilling knowledge of multiple domain-specific networks after co-aligning their features with the help of adapters and centered kernel alignment, and shows that the universal representations can be further refined for previously unseen domains by an efficient adaptation step.
- 113
- Synthesizing Informative Training Samples with GAN112
This paper proposes a novel method to synthesize Informative Training samples with GAN (IT-GAN), which freezes a pre-trained GAN model and learns the informative latent vectors that correspond to informative training samples.
- 100
- Knowledge Distillation for Multi-task Learning95
This work proposes a knowledge distillation based method that can optimize a multi-task learning model in a more balanced way and achieve better overall performance.
- 88
- Unsupervised Learning of Landmarks by Descriptor Vector Exchange71
A new perspective on the equivariance approach is developed by noting that dense landmark detectors can be interpreted as local image descriptors equipped with invariance to intra-category variations, and proposing a direct method to enforce such an invariance in the standard equivariant loss.
- Learning Multiple Dense Prediction Tasks from Partially Annotated Data64
This paper proposes a multi-task training procedure that successfully leverages task relations to supervise its multi- task learning when data is partially annotated, and learns to map each task pair to a joint pairwise task-space which enables sharing information between them in a computationally efficient way through another network conditioned on task pairs.
- Continual Representation Learning for Biometric Identification55
This paper proposes a new continual learning setting, namely "continual representation learning", which focuses on learning better representation in a continuous way, and demonstrates that existing CL methods can improve the representation in the new setting.
- 50
- Few‐shot Learning of Homogeneous Human Locomotion Styles49
This paper proposes a transfer learning approach for adapting a learned neural network to characters that move in different styles from those on which the original neural network is trained, and introduces a canonical polyadic tensor decomposition to reduce the amount of parameters required for learning from each new style.
- Improving Semantic Correspondence with Viewpoint-Guided Spherical Maps48
This work proposes a new semantic correspondence estimation method that supplements state-of-the-art self-supervised features with 3D understanding via a weak geometric spherical prior and provides a simple and effective way of injecting informative geometric priors into the learned representation while requiring only weak viewpoint information.
- Universal Representations: A Unified Look at Multiple Task and Domain Learning47
This work rigorously shows that universal representations achieve state-of-the-art performances in learning of multiple dense prediction problems in NYU-v2 and Cityscapes, multiple image classification problems from diverse domains in Visual Decathlon Dataset and cross-domain few-shot learning in MetaDataset.
- Not All Relations are Equal: Mining Informative Labels for Scene Graph Generation39
This work proposes a novel framework for SGG training that exploits relation labels based on their informativeness and shows that this approach can be used in conjunction with state-of-the-art SGG methods and improves their performance significantly in multiple metrics on the standard Visual Genome benchmark.
- Early Stuck Pipe Sign Detection with Depth-Domain 3D Convolutional Neural Network Using Actual Drilling Data39
A real-time stuck pipe prediction using the deep-learning approach with 3D-convolutional neural network approach with depth-domain data clip is studied, providing physical explanation of the black box model.
- InstanSeg: an embedding-based instance segmentation algorithm optimized for accurate, efficient and portable cell segmentation38
InstanSeg is a novel embedding-based instance segmentation pipeline designed to identify cells and nuclei in microscopy images that can significantly improve accuracy when compared to the most widely used alternative methods, while reducing the processing time by at least 60%.
- Early sign detection for the stuck pipe scenarios using unsupervised deep learning37
A neural network based unsupervised learning tool called Autoencoder is adapted for anticipating the ‘stuck’ events during the drilling process and Autoencoders on Recurrent Neural Networks are built to model the normal drilling activity, thereby detecting the stuck incidents as anomalous activity.
- A novel channel invariant architecture for the segmentation of cells and nuclei in multiplexed images using InstanSeg28
Channel Net is presented, a novel deep learning architecture for generating three-channel representations of multiplexed images irrespective of the number or ordering of imaged biomarkers and sets a new benchmark for the segmentation of cells and nuclei on public multiplexed imaging datasets.
- 28
- SpectraGAN27
SpectraGAN is a novel deep generative model that, upon training with real-world network traffic measurements, can produce high-fidelity synthetic mobile traffic data for new, arbitrary sized geographical regions over long periods.
- Object and Action Classification with Latent Window Parameters27
A generic framework to incorporate unobserved auxiliary information for classifying objects and actions is proposed, which allows us to automatically select a bounding box and its quadrants from which best to extract features.
- Unsupervised object learning from dense equivariant image labelling26
A new approach is proposed that, given a large number of images of an object and no other supervision, can extract a dense object-centric coordinate frame that is invariant to deformations of the images and comes with a dense equivariant labelling neural network that can map image pixels to their corresponding object coordinates.
- DepthCues: Evaluating Monocular Depth Perception in Large Vision Models25
This work introduces a new benchmark, DepthCues, designed to evaluate depth cue understanding, and presents findings across 20 diverse and representative pre-trained vision models, showing that human-like depth cues emerge in more recent larger models.
- Weakly Supervised Gaussian Networks for Action Detection25
A novel method is proposed, called WSGN, that learns to detect actions from weak supervision, using only video-level labels, that leads to significant gains in action detection for two standard benchmarks THU-MOS14 and Charades.
- Object and Action Classification with Latent Variables23
A method for incorporating latent variables into object and action classification and an exploration of a way to learn a better classifier by iterative expansion of the latent parameter space are provided.
- Conditional Image Generation for Learning the Structure of Visual Objects.21
It is demonstrated that this method can learn landmarks from synthetic image deformations or videos, all without manual supervision, while outperforming state-of-the-art unsupervised landmark detectors.
- Micromanipulation Using a Microassembly Workstation with Vision and Force Sensing21
A micromanipulation experiment that aims to locate the microspheres to the predefined configuration by using an integrated vision and force control scheme is successfully demonstrated to show the validity of the proposed methods.
- Learning Foreground-Background Segmentation from Improved Layered GANs20
A method to automatically synthesize paired photo-realistic images and segmentation masks for the use of training a foreground-background segmentation network that decomposes an image into foreground and background layers and avoids trivial decompositions is proposed.
- Developing robust vision modules for microsystems applications19
In this work, several robust vision modules are developed and implemented for fully automated micromanipulation, including autofocusing, object and end-effector detection, real-time tracking and optical system calibration modules.
- Looking 3D: Anomaly Detection with 2D-3D Alignment17
A novel transformer-based approach is proposed that explicitly learns the correspondence between the query image and reference 3D shape via feature alignment and leverages a customized attention mechanism for anomaly detection.
- Image Deconvolution with Deep Image and Kernel Priors17
It is shown that DIKP improve the performance of learning-free image deconvolution, and this is experimentally demonstrated on the standard benchmark of six standard test images in terms of PSNR and visual effects.
- 15
- DD-Ranking: Rethinking the Evaluation of Dataset Distillation14
DD-Ranking, a unified evaluation framework, along with new general evaluation metrics to uncover the true performance improvements achieved by different methods are proposed, which provide a more comprehensive and fair evaluation standard for future research advancements.
- Accelerating Self-Supervised Learning via Efficient Training Strategies14
This paper investigates reducing the training time of recent self-supervised methods by various model-agnostic strategies that have not been used for this problem, and studies three strategies: an extendable cyclic learning rate schedule, a matching progressive augmentation magnitude and image resolutions schedule, and a hard positive mining strategy based on augmentation difficulty.
- Explicit Neural Surfaces: Learning Continuous Geometry With Deformation Fields14
This work introduces Explicit Neural Surfaces (ENS), an efficient smooth surface representation that directly encodes topology with a deformation field from a known base domain and applies this representation to reconstruct explicit surfaces from multiple views.
- Modelling and unsupervised learning of symmetric deformable object categories14
It is shown that, if only raw images are given, it is possible to look instead for symmetries in the space of object deformations and provide an explanation of the ambiguities that arise in recovering the pose of symmetric objects from their shape or images and a way of discounting such ambigUities in learning.
- 13
- FaceMAE: Privacy-Preserving Face Recognition via Masked Autoencoders12
A novel framework FaceMAE, where the face privacy and recognition performance are considered simultaneously, is proposed, which consistently reduces at least 50% error rate on LFW, CFP-FP and AgeDB.
- 12
- 12
- Learning Action Changes by Measuring Verb-Adverb Textual Relationships11
The goal of this work is to understand the way actions are performed in videos, and to predict an adverb indicating a modification applied to the action, by measuring textual relationships between verbs and adverbs to generate a regression target representing the action change the authors aim to learn.
- Multi-task Learning with 3D-Aware Regularization11
This work proposes to address the problem of cross-task correlations in deep neural networks by introducing a structured 3D-aware regularizer which interfaces multiple tasks through the projection of features extracted from an image encoder to a shared 3D feature space and decodes them into their task output space through differentiable rendering.
- Who are you referring to? Coreference resolution in image narrations11
This paper proposes a new technique that learns to identify coref-erence chains using weak supervision, only from image-text pairs and a regularization using prior linguistic knowledge, and yields large performance gains over several strong baselines in resolving coreferences.
- An Unsupervised Learning Model for Pipe Stuck Predictions Using a Long Short-Term Memory Autoencoder Architecture11
This study has developed a prediction technique based on artificial intelligence built using an encoder-decoder, long short-term memory architecture that found that reconstruction errors increased prior to the pipe sticking in some cases and were sensitive to large variations in the drilling parameters.
- 11
- RECANTFormer: Referring Expression Comprehension with Varying Numbers of Targets10
RecantFormer is introduced, a one-stage method for GREC that combines a decoder-free (encoder-only) transformer architecture with DETR-like Hungarian matching that consistently outperforms baselines by significant margins in three GREC datasets.
- Learning Landmarks from Unaligned Data using Image Translation10
Modifications of the landmark detection model are shown to improve the quality of the learned detector leading to state-of-the-art unsupervised landmark detection performance in a number of challenging human pose and facial landmark detection benchmarks.
- 10
- 9
- 9
- 9
- ViewNet: Unsupervised Viewpoint Estimation from Conditional Generation8
This paper forms this as a self-supervised learning task, where image reconstruction provides the supervision needed to predict the camera viewpoint, and demonstrates that using a perspective spatial transformer allows efficient viewpoint learning, outperforming existing unsupervised approaches on synthetic data, and obtains competitive results on the challenging PASCAL3D+.
- Unsupervised Foreground-Background Segmentation with Equivariant Layered GANs.8
An unsupervised foreground-background segmentation method via training a segmentation network on the synthetic pseudo segmentation dataset generated from GANs, which are trained from a collection of images without annotations to explicitly disentangle foreground and background.
- 8
- Semi-supervised Viewpoint Estimation with Geometry-Aware Conditional Generation8
This paper proposes a semi-supervised viewpoint estimation method that can learn to infer viewpoint information from unlabeled image pairs, where two images differ by a viewpoint change.
- 8
- 8
- 8
- Learning to Annotate Part Segmentation with Gradient Matching7
This paper focuses on tackling semi-supervised part segmentation tasks by generating high-quality images with a pre-trained GAN and labelling the generated images with an automatic annotator, and formulate the annotator learning as a learning-to-learn problem.
- Attempt of Early Stuck Detection Using Unsupervised Deep Learning With Probability Mixture Model7
Unsupervised machine learning using data on the normal activities is a possible advanced method for early stuck detection, which is adopted in this study as a countermeasure to another concern that even normal activities involve various operations.
- Injecting Prior Knowledge into Image Caption Generation7
This paper proposes to improve the performance of the state-of-the-art image captioning models by incorporating two sources of prior knowledge: a conditional latent topic attention, that uses a set of latent variables (topics) as an anchor to generate highly probable words and a regularization technique that exploits the inductive biases in syntactic and semantic structure of captions and improves the generalization of image Captioning models.
- 7
- Early Stuck Detection Using Supervised and Unsupervised Machine Learning Approaches6
The results demonstrate that the unsupervised machine learning approach presents a better early stuck pipe detection capability and the proposed machine learning algorithm will be further improved in the future and the prediction result will be validated through actual operation.
- Visually Interpretable Subtask Reasoning for Visual Question Answering5
This work introduces VISTAR (Visually Interpretable Subtask-Aware Reasoning Model), a subtask-driven training framework that enhances both interpretability and reasoning by generating textual and visual explanations within MLLMs.
- 5
- 5
- 5
- Odd-One-Out: Anomaly Detection by Comparing with Neighbors4
This paper introduces a novel anomaly detection problem aimed at identifying ‘odd-looking’ objects within a scene by comparing them to other objects present and provides a comprehensive analysis of the method quantitatively and qualitatively on these benchmarks.
- 4
- 4
- Classification with Global, Local and Shared Features4
A framework that jointly learns and then uses multiple image windows for improved classification that is suited for multiple visual tasks such as classifying objects, scenes and actions is presented.
- Spatially-Adaptive Hash Encodings for Neural Surface Reconstruction3
This work proposes a learned approach which allows the network to choose its encoding basis as a function of space, by masking the contribution of features stored at separate grid resolutions, which allows the network to fit a wider range of frequencies without introducing noise.
- 3
- ViewNeRF: Unsupervised Viewpoint Estimation Using Category-Level Neural Radiance Fields3
ViewNeRF is introduced, a Neural Radiance Field-based viewpoint estimation method that learns to predict category-level viewpoints directly from images during training, to produce self-supervised reconstructions for whole object categories.
- 2
- 2
- 2
- 2
- Interactive Anomaly Detection for Articulated Objects via Motion Anticipation1
This paper introduces a tailored solution that detects functional anomalies by integrating vision, interaction, and anticipation, and introduces a new benchmark, PartNet-IAD, for interactive AD, which includes articulated objects with realistic functional anomalies.
- Supervised and unsupervised deep learning-based approaches for studying DNA replication spatiotemporal dynamics1
It is demonstrated that a convolutional neural network trained to classify S-phase stages based on DAPI and EdU patterns could identify altered replication dynamics in Rif1-deficient mouse embryonic stem cells (mESCs), revealing a skewed distribution across the various S-phase stages.
- Learning to Predict Keypoints and Structure of Articulated Objects without Supervision1
A method that learns to infer the skeleton structure of a novel articulated object from a single image, in terms of joints and rigid links connecting them, without supervision is presented.
- Novel estimation and control techniques in micromanipulation using vision and force feedback1
A fully automated dexterous micromanipulation strategy based on vision and force feedback is developed and a robust vision based control architecture is proposed and implemented to compensate errors due to the uncertainties about the position, behavior and shape of the micro objects to be manipulated.
- Underwater Object Detection for Autonomous Observation and Sampling Using General-Purpose Object Detection Techniques–
This study evaluated the performance of various object detection methods across an open underwater image dataset, focusing on detecting aquatic animals for performance evaluation because they were the primary focus of publicly available datasets.
- –
- –
- –
- HumMorph: Generalized Dynamic Human Neural Fields From Few Views–
The experimental results demonstrate that the HumMorph architecture is more robust to errors in the noisy parameters and clearly outper-forms the state of the art in this setting.
- –
- Coarse or Fine? Recognising Action End States without Labels–
This work focuses on the problem of recognising the end state of an action in an image, which is critical for understanding what action is performed and in which manner, and proposes an augmentation method to synthesise training data.
- Distilling Representations from GAN Generator via Squeeze and Span–
This paper squeezes the generator features into representations that are invariant to semantic-preserving transformations through a network before they are distilled into the student network and spans the distilled representation of the synthetic domain to the real domain.
- –
- Learning to Caption Images with Two-Stream Attention and Sentence Auto-Encoder–
A two-stream attention mechanism that can automatically discover latent categories and relate them to image regions based on the previously generated words and a regularization technique that encapsulates the syntactic and semantic structure of captions and improves the optimization of the image captioning model are proposed.
- –
- –
- Model-based vs. model-free visual servoing: a performance evaluation in microsystems–
In this paper, model-based and model-free image based visual servoing approaches are implemented on a microassembly workstation, and their regulation and tracking performances are evaluated.
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 2026-10-12. 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.