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
- Stefano SoattoSuggested from co-authorship
- Andrew ZissermanSuggested from co-authorship
Is this you? Claim this profile to confirm or dismiss it.
Works319 from public data
- Deep Face Recognition5,690
It is shown how a very large scale dataset can be assembled by a combination of automation and human in the loop, and the trade off between data purity and time is discussed.
- Fully-Convolutional Siamese Networks for Object Tracking4,457
A basic tracking algorithm is equipped with a novel fully-convolutional Siamese network trained end-to-end on the ILSVRC15 dataset for object detection in video and achieves state-of-the-art performance in multiple benchmarks.
- Deep Image Prior3,961
It is shown that a randomly-initialized neural network can be used as a handcrafted prior with excellent results in standard inverse problems such as denoising, super-resolution, and inpainting.
- Describing Textures in the Wild3,857
This work identifies a vocabulary of forty-seven texture terms and uses them to describe a large dataset of patterns collected "in the wild", and shows that they both outperform specialized texture descriptors not only on this problem, but also in established material recognition datasets.
- Vlfeat3,535
VLFeat is an open and portable library of computer vision algorithms that includes rigorous implementations of common building blocks such as feature detectors, feature extractors, (hierarchical) k-means clustering, randomized kd-tree matching, and super-pixelization.
- Return of the Devil in the Details: Delving Deep into Convolutional Nets3,509
It is shown that the data augmentation techniques commonly applied to CNN-based methods can also be applied to shallow methods, and result in an analogous performance boost, and it is identified that the dimensionality of the CNN output layer can be reduced significantly without having an adverse effect on performance.
- MatConvNet2,984
MatConvNet exposes the building blocks of CNNs as easy-to-use MATLAB functions, providing routines for computing convolutions with filter banks, feature pooling, normalisation, and much more.
- 2,774
- Understanding deep image representations by inverting them2,127
This paper conducts a direct analysis of the visual information contained in representations by asking the following question: given an encoding of an image, to which extent is it possible to reconstruct the image itself?
- VGGT: Visual Geometry Grounded Transformer1,963
VGGT is presented, a feed-forward neural network that directly infers all key 3D attributes of a scene, including camera parameters, point maps, depth maps, and 3D point tracks, from one, a few, or hundreds of its views, a step forward in 3D computer vision.
- Synthetic Data for Text Localisation in Natural Images1,584
The relation of FCRN to the recently-introduced YOLO detector, as well as other end-to-end object detection systems based on deep learning, are discussed.
- Speeding up Convolutional Neural Networks with Low Rank Expansions1,583
Two simple schemes for drastically speeding up convolutional neural networks are presented, achieved by exploiting cross-channel or filter redundancy to construct a low rank basis of filters that are rank-1 in the spatial domain.
- 1,519
- End-to-End Representation Learning for Correlation Filter Based Tracking1,469
This work is the first to overcome this limitation by interpreting the Correlation Filter learner, which has a closed-form solution, as a differentiable layer in a deep neural network, which enables learning deep features that are tightly coupled to the Cor correlation filter.
- Reading Text in the Wild with Convolutional Neural Networks1,189
An end-to-end system for text spotting—localising and recognising text in natural scene images—and text based image retrieval and a real-world application to allow thousands of hours of news footage to be instantly searchable via a text query is demonstrated.
- 1,084
- 1,054
- Invariant Information Clustering for Unsupervised Image Classification and Segmentation1,005
A novel clustering objective is presented that learns a neural network classifier from scratch, given only unlabelled data samples, and discovers clusters that accurately match semantic classes, achieving state-of-the-art results in eight unsupervised clustering benchmarks spanning image classification and segmentation.
- Quick Shift and Kernel Methods for Mode Seeking1,005
It is shown that the complexity of the recently introduced medoid-shift algorithm in clustering N points is O(N 2), with a small constant, if the underlying distance is Euclidean, which makes medoid shift considerably faster than mean shift, contrarily to what previously believed.
- Vggsound: A Large-Scale Audio-Visual Dataset992
The goal is to collect a large-scale audio-visual dataset with low label noise from videos ‘in the wild’ using computer vision techniques and investigates various Convolutional Neural Network architectures and aggregation approaches to establish audio recognition baselines for this new dataset.
- The devil is in the details: an evaluation of recent feature encoding methods954
A rigorous evaluation of novel encodings for bag of visual words models by identifying both those aspects of each method which are particularly important to achieve good performance, and those aspects which are less critical, which allows a consistent comparative analysis of these encoding methods.
- HPatches: A Benchmark and Evaluation of Handcrafted and Learned Local Descriptors913
A novel benchmark for evaluating local image descriptors is proposed and it is shown that a simple normalisation of traditional hand-crafted descriptors can boost their performance to the level of deep learning based descriptors within a realistic benchmarks evaluation.
- 882
- Improved Texture Networks: Maximizing Quality and Diversity in Feed-Forward Stylization and Texture Synthesis877
This work introduces an instance normalization module to replace batch normalization with significant improvements to the quality of image stylization and improves diversity by introducing a new learning formulation that encourages generators to sample unbiasedly from the Julesz texture ensemble.
- 872
- 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.
- The Sixth Visual Object Tracking VOT2018 Challenge Results804
The Visual Object Tracking challenge VOT2018 is the sixth annual tracker benchmarking activity organized by the VOT initiative; results of over eighty trackers are presented; many are state-of-the-art trackers published at major computer vision conferences or in journals in the recent years.
- The Visual Object Tracking VOT2016 Challenge Results776
The Visual Object Tracking challenge VOT2016 goes beyond its predecessors by introducing a new semi-automatic ground truth bounding box annotation methodology and extending the evaluation system with the no-reset experiment.
- 773
- 726
- 718
- Deep Features for Text Spotting694
A Convolutional Neural Network classifier is developed that can be used for text spotting in natural images and a method of automated data mining of Flickr, that generates word and character level annotations is used to form an end-to-end, state-of-the-art text spotting system.
- 677
- Understanding Image Representations by Measuring Their Equivariance and Equivalence654
This work investigates two key mathematical properties of representations: equivariance and equivalence and identifies several predictors of geometric and architectural compatibility, including the spatial resolution of the representation and the complexity and depth of the models.
- 578
- 571
- 531
- Accurate brain age prediction with lightweight deep neural networks514
A lightweight deep learning model, Simple Fully Convolutional Network (SFCN), is presented, achieving state-of-the-art brain age prediction and sex classification performance in UK Biobank MRI brain imaging data and performs better than widely-used regression models.
- 514
- Understanding Deep Networks via Extremal Perturbations and Smooth Masks511
Some of the shortcomings of existing approaches to perturbation analysis are discussed and the concept of extremal perturbations are introduced, which are theoretically grounded and interpretable and allow us to remove all tunable weighing factors from the optimization problem.
- Learning feed-forward one-shot learners500
This paper constructs the learner as a second deep network, called a learnet, which predicts the parameters of a pupil network from a single exemplar, and obtains an efficient feed-forward one-shot learner, trained end-to-end by minimizing a one- shot classification objective in a learning to learn formulation.
- Visualizing Deep Convolutional Neural Networks Using Natural Pre-images496
This paper studies several landmark representations, both shallow and deep, by a number of complementary visualization techniques based on the concept of “natural pre-image”, and shows that several layers in CNNs retain photographically accurate information about the image, with different degrees of geometric and photometric invariance.
- Fisher Vector Faces in the Wild487
This paper shows that Fisher vectors on densely sampled SIFT features are capable of achieving state-of-the-art face verification performance on the challenging “Labeled Faces in the Wild” benchmark, and shows that a compact descriptor can be learnt from them using discriminative metric learning.
- Splatter Image: Ultra-Fast Single-View 3D Reconstruction468
The Splatter Image is introduced, an ultra-efficient approach for monocular 3D object reconstruction based on Gaussian Splatting that achieves better results in terms of PSNR, LPIPS, and other metries while training and evaluating much faster than prior works.
- 462
- 434
- 429
- Learning to Discover Novel Visual Categories via Deep Transfer Clustering409
The problem of discovering novel object categories in an image collection is considered, and Deep Embedded Clustering is extended to a transfer learning setting, and the algorithm is improved by introducing a representation bottleneck, temporal ensembling, and consistency.
- 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.
- Training-Free Layout Control with Cross-Attention Guidance387
This work proposes a simple approach that achieves robust layout control without the need for training or fine-tuning of the image generator, and demonstrates the versatility of layout guidance by extending it to applications such as editing the layout and context of real images.
- RealFusion 360° Reconstruction of Any Object from a Single Image385
This work takes an off-the-self conditional image generator based on diffusion and engineer a prompt that encourages it to “dream up” novel views of the object, and fuse the given input view, the conditional prior, and other regularizers into a final, consistent reconstruction.
- Deep Filter Banks for Texture Recognition, Description, and Segmentation374
A human-interpretable vocabulary of texture attributes to describe common texture patterns is proposed, complemented by a new describable texture dataset for benchmarking and state-of-the-art performance in numerous datasets well beyond textures are obtained.
- 362
- Generalized Category Discovery356
This paper establishes strong baselines by taking state-of-the-art algorithms from novel category discovery and adapting them for this task, and introduces a simple yet effective semi-supervised k-means method to cluster the unlabelled data into seen and unseen classes automatically, substantially outperforming the baselines.
- 353
- Unsupervised Learning of Probably Symmetric Deformable 3D Objects From Images in the Wild341
This paper proposes a method to learn 3D deformable object categories from raw single-view images, without external supervision, based on an autoencoder that factors each input image into depth, albedo, viewpoint and illumination.
- What does CLIP know about a red circle? Visual prompt engineering for VLMs303
This work explores the idea of visual prompt engineering for solving computer vision tasks beyond classification by editing in image space instead of text, and discovers an emergent ability of CLIP, where, by simply drawing a red circle around an object, it can direct the model’s attention to that region, while also maintaining global information.
- 302
- Net2Vec: Quantifying and Explaining How Concepts are Encoded by Filters in Deep Neural Networks301
The Net2Vec framework is introduced, in which semantic concepts are mapped to vectorial embeddings based on corresponding filter responses, and it is shown that in most cases, multiple filters are required to code for a concept, and that filterembeddings are able to better characterize the meaning of a representation and its relationship to other concepts.
- Emotion Recognition in Speech using Cross-Modal Transfer in the Wild298
A strong teacher network for facial emotion recognition that achieves the state of the art on a standard benchmark is developed and it is shown that the speech emotion embedding can be used for speech emotion recognition on external benchmark datasets.
- BANMo: Building Animatable 3D Neural Models from Many Casual Videos260
This work aims to create high-fidelity, articulated 3D models from many casual RGB videos in a differentiable rendering framework, and introduces neural blend skinning models that allow for differentiable and invertible articulated deformations.
- Neural Feature Fusion Fields: 3D Distillation of Self-Supervised 2D Image Representations257
A method that improves dense image feature extractors when the latter are applied to the analysis of multiple images reconstructible as a $3D$ scene and consistently improves over the self-supervised $2D$ baselines.
- Localizing Visual Sounds the Hard Way252
The key technical contribution is to show that, by training the network to explicitly discriminate challenging image fragments, even for images that do contain the object emitting the sound, the authors can significantly boost the localization performance by introducing a mechanism to mine hard samples and add them to a contrastive learning formulation automatically.
- 251
- AutoNovel: Automatically Discovering and Learning Novel Visual Categories220
This work suggests that the common approach of bootstrapping an image representation using the labelled data only introduces an unwanted bias, and that this can be avoided by using self-supervised learning to train the representation from scratch on the union of labelled and unlabelled data.
- 215
- Deep Fisher Networks for Large-Scale Image Classification212
This paper proposes a version of the state-of-the-art Fisher vector image encoding that can be stacked in multiple layers, and significantly improves on standard Fisher vectors, and obtains competitive results with deep convolutional networks at a smaller computational learning cost.
- Deep Spectral Methods: A Surprisingly Strong Baseline for Unsupervised Semantic Segmentation and Localization210
Experiments on complex datasets demonstrate that the simple spectral method outperforms the state-of-the-art in unsupervised localization and segmentation by a significant margin, and can be readily used for a variety of complex image editing tasks, such as background removal and compositing.
- 209
- 200
- Multi-modal Self-Supervision from Generalized Data Transformations197
The framework of Generalized Data Transformations is introduced to reduce several recent self-supervised learning objectives to a single formulation for ease of comparison, analysis, and extension, and allow a choice between being invariant or distinctive to data transformations, obtaining different supervisory signals, and derive the conditions that combinations of transformations must obey in order to lead to well-posed learning objectives.
- Localizing Objects with Smart Dictionaries192
This work presents an approach to determine the category and location of objects in images that reduces the size of a large generic dictionary to the low hundreds while increasing classification performance compared to k-means, and performs feature-based categorization efficiently on a dense grid.
- Exemplar Fine-Tuning for 3D Human Model Fitting Towards In-the-Wild 3D Human Pose Estimation186
This paper augments existing 2D datasets with high-quality 3D pose fits by augmenting them with Exemplar Fine-Tuning (EFT), and shows that EFT produces 3D annotations that result in better downstream performance and are qualitatively preferable in an extensive human-based assessment.
- 186
- Long-Term Tracking in the Wild: A Benchmark184
The OxUvA dataset and benchmark for evaluating single-object tracking algorithms is introduced, offering the community a large and diverse benchmark to enable the design and evaluation of tracking methods ready to be used “in the wild”.
- 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.
- DynamicStereo: Consistent Dynamic Depth from Stereo Videos172
This work proposes DynamicStereo, a novel transformer-based architecture to estimate dispar-ity for stereo videos, which learns to pool information from neighboring frames to improve the temporal consistency of its predictions.
- Pylon Model for Semantic Segmentation163
This paper shows that if, instead of a flat partitioning, the image is represented by a hierarchical segmentation tree, then the resulting energy combining unary and boundary terms can still be optimized using graph cut (with all the corresponding benefits of global optimality and efficiency).
- Supervising the New with the Old: Learning SFM from SFM156
This paper proposes a probabilistic learning formulation where the network predicts distributions over variables rather than specific values, and shows that this formulation can learn and account for the defects of SFM, helping to integrate different sources of information and boosting the overall performance of the network.
- Learning Covariant Feature Detectors148
This paper proposes to cast detection as a regression problem, enabling the use of powerful regressors such as deep neural networks, and derives a covariance constraint that can be used to automatically learn which visual structures provide stable anchors for local feature detection.
- Flash3D: Feed-Forward Generalisable 3D Scene Reconstruction from a Single Image145
This work proposes Flash3D, a method for scene reconstruction and novel view synthesis from a single image which is both very generalisable and efficient, and achieves state-of-the-art results when trained and tested on RealEstate10k.
- 145
- HOLODIFFUSION: Training a 3D Diffusion Model Using 2D Images142
This work introduces a new diffusion setup that can be trained, end-to-end, with only posed 2D images for supervision, and proposes an image formation model that decouples model memory from spatial memory that is competitive in terms of sample quality and fidelity to existing approaches for 3D generative modeling.
- 142
- PC2: Projection-Conditioned Point Cloud Diffusion for Single-Image 3D Reconstruction141
This paper proposes a novel method for single-image 3D reconstruction which generates a sparse point cloud via a conditional denoising diffusion process which not only performs well on synthetic benchmarks but also gives large qualitative improvements on complex real-world data.
- It Takes (Only) Two: Adversarial Generator-Encoder Networks141
A new autoencoder-type architecture that is trainable in an unsupervised mode, sustains both generation and inference, and has the quality of conditional and unconditional samples boosted by adversarial learning is presented.
- Viewset Diffusion: (0-)Image-Conditioned 3D Generative Models from 2D Data137
Viewset Diffusion is presented, a diffusion-based generator that outputs 3D objects while only using multi-view 2D data for supervision, and is trained using only rendering losses using as few as three views per viewset.
- MagicPony: Learning Articulated 3D Animals in the Wild135
A new method is presented, dubbed MagicPony, that learns this predictor purely from in-the-wild single-view images of the object category, with minimal assumptions about the topology of deformation, to overcome local optima in viewpoint estimation.
- Continual Detection Transformer for Incremental Object Detection134
This paper proposes a ContinuaL DEtection TRansformer (CL-DETR), a new method for transformer-based IOD which enables effective usage of KD and ER in this context and introduces a Detector Knowledge Distillation (DKD) loss, focusing on the most informative and reliable predictions from old versions of the model.
- GES: Generalized Exponential Splatting for Efficient Radiance Field Rendering133
GES (Generalized Exponential Splatting), a novel representation that employs Generalized Exponential Function (GEF) to model 3D scenes, requiring far fewer parti-cles to represent a scene and thus significantly outperforming Gaussian Splatting methods in efficiency with a plug-and-play replacement ability for Gaussian-based utilities.
- Salient Deconvolutional Networks133
A family of reversed networks is introduced that generalizes and relates deconvolution, backpropagation and network saliency, and is used to thoroughly investigate and compare these methods in terms of quality and meaning of the produced images, and of what architectural choices are important in determining these properties.
- 125
- ShapeStacks: Learning Vision-Based Physical Intuition for Generalised Object Stacking123
This paper provides ShapeStacks, a simulation-based dataset featuring 20,000 stack configurations composed of a variety of elementary geometric primitives richly annotated regarding semantics and structural stability, and trains visual classifiers for binary stability prediction on the data and scrutinise their learned physical intuition.
- 120
- 119
- VMem: Consistent Interactive Video Scene Generation with Surfel-Indexed View Memory117
Surfel-Indexed View Memory is introduced, a memory module that remembers past views by indexing them geometrically based on the 3D surface elements (surfels) they have observed, which enables efficient retrieval of the most relevant past views when generating new ones.
- NightOwls: A Pedestrians at Night Dataset108
A comprehensive public dataset, NightOwls, is introduced, for pedestrian detection at night, due to variable and low illumination, reflections, blur, and changing contrast in comparison to daytime conditions.
- NeuroMorph: Unsupervised Shape Interpolation and Correspondence in One Go107
A new neural network architecture that takes as input two 3D shapes and produces in one go, in a single feed forward pass, a smooth interpolation and point-to-point correspondences between them, matching or surpassing the performance of recent unsupervised and supervised methods on multiple benchmarks.
- R-CNN minus R97
This paper designs and evaluates a detector that uses a trivial region generation scheme, constant for each image, and results in an excellent and fast detector that does not require to process an image with algorithms other than the CNN itself.
- 96
- 95
- 94
- Relaxed Softmax: Efficient Confidence Auto-Calibration for Safe Pedestrian Detection94
This paper investigates the problem of learning in an end-to-end manner object detectors that are accurate while providing an unbiased estimate of the reliablity of their own predictions by proposing a modi-cation of the standard softmax layer where a probabilistic confidence score is explicitly pre-multiplied into the incoming activations to modulate confidence.
- 94
- Generalized RBF feature maps for Efficient Detection94
This paper completes the construction and combines the two techniques to obtain explicit feature maps for the generalized RBF kernels, and investigates a learning method using l 1 regularization to encourage sparsity in the final vector representation, and thus reduce its dimension.
- 88
- Semi-convolutional Operators for Instance Segmentation88
It is shown theoretically and empirically that constructing dense pixel embeddings that can separate object instances cannot be easily achieved using convolutional operators, and that simple modifications, which are called semi-convolutional, have a much better chance of succeeding at this task.
- 86
- Free3D: Consistent Novel View Synthesis Without 3D Representation85
Free3D is introduced, a simple accurate method for monocular open-set novel view synthesis (NVS) that achieves significant improvements without resorting to an explicit 3D representation, which is slow and memory-consuming, and without training an additional network for 3D reconstruction.
- Exemplar Fine-Tuning for 3D Human Pose Fitting Towards In-the-Wild 3D Human Pose Estimation85
This work proposes a method for building large collections of human poses with full 3D annotations captured `in the wild', for which specialized capture equipment cannot be used and uses EFT to augment these large in-the-wild datasets with plausible and accurate 3D pose annotations.
- Structured output regression for detection with partial truncation85
A structured output model for object category detection that explicitly accounts for alignment, multiple aspects and partial truncation in both training and inference is developed.
- DOVE: Learning Deformable 3D Objects by Watching Videos82
DOVE is a method that learns textured 3D models of deformable object categories from monocular videos available online, without keypoint, viewpoint or template shape supervision, and produces temporally consistent 3D models, which can be animated and rendered from arbitrary viewpoints.
- Simultaneous Object Detection and Ranking with Weak Supervision82
A discriminative learning approach is developed, which proposes a structured output formulation for weakly annotated images where full annotations are treated as latent variables and proposes to optimize a ranking objective function, allowing the method to more effectively use negatively labeled images to improve detection average precision performance.
- Online Clustered Codebook79
This paper presents a simple alternative method for online codebook learning, Clustering VQ-VAE (CVQ-VAE), which selects encoded features as anchors to update the "dead" codevectors, while optimising the codebooks which are alive via the original loss.
- Diffusion Models for Open-Vocabulary Segmentation76
OVDiff is presented, a novel method that leverages generative text-to-image diffusion models for unsupervised open-vocabulary segmentation and shows strong performance on a range of benchmarks, obtaining a lead of more than 5% over prior work on PASCAL VOC.
- 76
- Self-Supervised Learning of Geometrically Stable Features Through Probabilistic Introspection73
This paper shows empirically that a network pre- trained in this manner requires significantly less supervision to learn semantic object parts compared to numerous pre-training alternatives, and shows that the pre-trained representation is excellent for semantic object matching.
- 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.
- Descriptor Learning Using Convex Optimisation71
The objective of this work is to learn descriptors suitable for the sparse feature detectors used in viewpoint invariant matching, and it is shown that learning the pooling regions for the descriptor can be formulated as a convex optimisation problem selecting the regions using sparsity.
- AnchorNet: A Weakly Supervised Network to Learn Geometry-Sensitive Features for Semantic Matching70
This work proposes a deep network, termed AnchorNet, that produces image representations that are well-suited for semantic matching that are trained only with weak image-level labels and improves results of state-of-the-art semantic matching methods such as the Deformable Spatial Pyramid or the Proposal Flow methods.
- Invariant Information Distillation for Unsupervised Image Segmentation and Clustering69
A simple loss based on information theory is used to extract meaningful representations directly from raw images, making it the first end-to-end unsupervised learning method that learns densely and efficiently (i.e. without sampling) for semantic segmentation.
- Cross Pixel Optical-Flow Similarity for Self-supervised Learning68
This work uses motion cues in the form of optical flow, to supervise representations of static images, and achieves state-of-the-art results in self-supervision using motion cues, competitive results for self- supervision in general, and is overall state of the art inSelf-supervised pretraining for semantic image segmentation.
- 3D Convex Splatting: Radiance Field Rendering with 3D Smooth Convexes65
A novel method, named 3D Convex Splatting (3DCS), which leverages 3D smooth convexes as primitives for modeling geometrically-meaningful radiance fields from multi-view images, highlighting the potential of 3D Convex Splatting to become the new standard for high-quality scene reconstruction and novel view synthesis.
- NeuralDiff: Segmenting 3D objects that move in egocentric videos65
It is demonstrated that the method can successfully separate the different types of motion, outperforming recent neural rendering baselines at this task, and can accurately segment the moving objects.
- 63
- 62
- Amodal3R: Amodal 3D Reconstruction from Occluded 2D Images61
It is demonstrated that, by training solely on synthetic data, Amodal3R learns to recover full 3D objects even in the presence of occlusions in real scenes, thereby establishing a new benchmark for occlusion-aware 3D reconstruction.
- 60
- Triangle Splatting for Real-Time Radiance Field Rendering59
A differentiable renderer that directly optimizes triangles via end-to-end gradients is developed, rendering each triangle as differentiable splats, combining the efficiency of triangles with the adaptive density of representations based on independent primitives.
- Explanations for Attributing Deep Neural Network Predictions59
This chapter introduces Meta-Predictors as Explanations, a principled framework for learning explanations for any black box algorithm, and Meaningful Perturbations, an instantiation of the paradigm applied to the problem of attribution.
- Self-supervised object detection from audio-visual correspondence54
This work extracts a supervisory signal from audio-visual data, using the audio component to “teach” the object detector, and outperform previous unsupervised and weakly-supervised detectors for the task of object detection and sound source localization.
- On Compositions of Transformations in Contrastive Self-Supervised Learning54
It is found that being invariant to certain transformations and distinctive to others is critical to learning effective video representations, improving the state-of-the-art for multiple benchmarks by a large margin, and even surpassing supervised pretraining.
- 48
- 47
- 46
- Optimising a Simple Fully Convolutional Network for Accurate Brain Age Prediction in the PAC 2019 Challenge45
A lightweight deep convolutional neural network architecture, Simple Fully Convolutional Neural Network (SFCN) is applied, and several techniques including data augmentation, transfer learning, model ensemble, and bias correction are combined for brain age prediction.
- 45
- Fully-trainable deep matching43
This paper rewriting the complete DM algorithm as a convolutional neural network results in a novel deep architecture for image matching that involves a number of new layer types and that, similar to recent networks for image segmentation, has a U-topology.
- 41
- 39
- 38
- 38
- 38
- 36
- 36
- 35
- 34
- 34
- 30
- 29
- 28
- 28
- 28
- 27
- 26
- 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.
- 25
- 23
- 23
- 23
- 22
- 21
- 21
- 21
- 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.
- 20
- 20
- 20
- 20
- 20
- 20
- 20
- 19
- 19
- 18
- 18
- 17
- 17
- 17
- 17
- 16
- 16
- 16
- 16
- Flex3D: Feed-Forward 3D Generation with Flexible Reconstruction Model and Input View Curation15
Flex3D is a novel two-stage framework capable of leveraging an arbitrary number of high-quality input views, built upon a transformer architecture that can effectively process an arbitrary number of inputs, enabling efficient and detailed 3D generation.
- 15
- 15
- 15
- 14
- 14
- 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.
- 14
- 13
- 13
- 13
- 12
- 11
- 11
- 11
- 11
- 11
- 10
- 10
- 10
- 10
- 10
- 9
- 9
- 9
- 9
- 9
- 9
- 9
- Learning Landmarks from Unaligned Data using Image Translation9
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.
- 9
- 8
- 8
- 8
- 8
- 8
- 8
- 7
- 7
- 7
- 7
- 7
- 6
- 6
- 6
- 6
- 6
- 6
- 6
- 6
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- 5
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
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 2026-10-13. 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.