Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks131 from public data
- 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.
- Image Synthesis in Multi-Contrast MRI With Conditional Generative Adversarial Networks608
The proposed approach preserves intermediate-to-high frequency details via an adversarial loss, and it offers enhanced synthesis performance via pixel-wise and perceptual losses for registered multi-contrast images and a cycle-consistency loss for unregistered images.
- 544
- Visual saliency estimation by nonlinearly integrating features using region covariances417
This work proposes to use covariance matrices of simple image features (known as region covariance descriptors in the computer vision community) as meta-features for saliency estimation and demonstrates that the proposed approach outperforms the state-of-art models on various tasks including prediction of human eye fixations, salient object detection, and image-retargeting.
- Automatic Description Generation from Images: A Survey of Models, Datasets, and Evaluation Measures397
This survey classify the existing approaches based on how they conceptualize this problem, viz., models that cast description as either generation problem or as a retrieval problem over a visual or multimodal representational space.
- Structure-preserving image smoothing via region covariances219
This study proposes an alternative yet simple image smoothing approach which depends on covariance matrices of simple image features, aka the region covariances, and uses second order statistics as a patch descriptor to implicitly capture local structure and texture information.
- RecipeQA: A Challenge Dataset for Multimodal Comprehension of Cooking Recipes198
This work introduces RecipeQA, a dataset for multimodal comprehension of cooking recipes, a set of comprehension and reasoning tasks that require joint understanding of images and text, capturing the temporal flow of events and making sense of procedural knowledge.
- Re-evaluating Automatic Metrics for Image Captioning195
This paper provides an in-depth evaluation of the existing image captioning metrics through a series of carefully designed experiments and explores the utilization of the recently proposed Word Mover’s Distance document metric for the purpose of image Captioning.
- mustGAN: multi-stream Generative Adversarial Networks for MR Image Synthesis185
Quantitative and radiological assessments on T1, T2, PD-weighted, and FLAIR images clearly demonstrate the superior performance of the proposed multi-stream approach that aggregates information across multiple source images via a mixture of multiple one- to-one streams and a joint many-to-one stream.
- 176
- The State of the Art in HDR Deghosting: A Survey and Evaluation127
A taxonomy of deghosting algorithms is proposed which can be used to group existing and future algorithms into meaningful classes, and the results of a subjective experiment are shared which aims to evaluate various state‐of‐the‐art de ghosting algorithms.
- An Objective Deghosting Quality Metric for HDR Images113
An objective metric is proposed which aims to simplify the process of reconstructing high dynamic range images of a complex scene involving moving objects and dynamic backgrounds by taking a stack of input exposures and the deghosting result and producing a set of artifact maps for different types of artifacts.
- Inst-Inpaint: Instructing to Remove Objects with Diffusion Models93
This work presents a novel inpainting framework, Inst-Inpaint, that can remove objects from images based on the instructions given as text prompts and compares methods with different evaluation metrics that measure the quality and accuracy of the models and show significant quantitative and qualitative improvements.
- Disconnected Skeleton: Shape at Its Absolute Scale88
This work presents a new skeletal representation along with a matching framework to address the deformable shape recognition problem and replaces the local coordinate frame with a global Euclidean frame supported by additional mechanisms to handle articulations and local boundary deformations.
- 87
- Deformable part-based tracking by coupled global and local correlation filters81
A deformable part-based correlation filter tracking approach which depends on coupled interactions between a global filter and several part filters, which gives significantly better results compared with the state-of-the-art trackers.
- 78
- HyperE2VID: Improving Event-Based Video Reconstruction via Hypernetworks74
This study proposes HyperE2VID, a dynamic neural network architecture for event-based video reconstruction that uses hypernetworks to generate per-pixel adaptive filters guided by a context fusion module that combines information from event voxel grids and previously reconstructed intensity images.
- Neural Natural Language Generation: A Survey on Multilinguality, Multimodality, Controllability and Learning68
This state-of-the-art report investigates the recent developments and applications of NNLG in its full extent from a multidimensional view, covering critical perspectives such as multimodality, multilinguality, controllability and learning strategies.
- The Thermal Infrared Visual Object Tracking VOT-TIR2016 Challenge Results66
A significant general improvement of results has been observed, which partly compensate for the more difficult sequences introduced into the dataset in VOT-TIR2016.
- Computer vision based mouse66
A computer vision based mouse, which can control and command the cursor of a computer or a computerized system using a camera, which has regions corresponding to buttons for clicking.
- Cross-lingual Visual Pre-training for Multimodal Machine Translation57
This paper extends the translation language modelling with masked region classification and performs pre-training with three-way parallel vision & language corpora to learn visually-grounded cross-lingual representations and shows that when fine-tuned for multimodal machine translation, these models obtain state-of-the-art performance.
- Graph Transduction as a Non-Cooperative Game55
This article proposes a radically different perspective on transduction, based on game-theoretic notions, that is inherently a multiclass approach and imposes no constraint whatsoever on the structure of the pairwise similarity matrix, being able to naturally deal with asymmetric and negative similarities alike.
- CRAFT: A Benchmark for Causal Reasoning About Forces and inTeractions52
The results show that even though the questions in CRAFT are easy for humans, the tested baseline models, including existing state-of-the-art methods, do not yet deal with the challenges posed in the benchmark.
- SLAMP: Stochastic Latent Appearance and Motion Prediction50
This work reason about appearance and motion in the video stochastically by predicting the future based on the motion history, and significantly outperforms state-of-the-art models on two challenging real-world autonomous driving datasets with complex motion and dynamic background.
- Dissimilarity between two skeletal trees in a context48
A new tree-edit based shape matching method which uses a recent coarse skeleton representation which allows both shapes and shape categories in the form of depth-1 trees to easily integrate the influence of the categories into shape dissimilarity measurements.
- Hyperspectral image denoising via self-modulating convolutional neural networks43
A self-modulating convolutional neural network which is referred to as SM-CNN is introduced, which utilizes correlated spectral and spatial information to transform the features in an adaptive manner based on the adjacent spectral data, enhancing the network's ability to handle complex noise.
- 42
- A Distributed Representation Based Query Expansion Approach for Image Captioning38
The core idea of the method is to translate the given visual query into a distributional semantics based form, which is generated by the average of the sentence vectors extracted from the captions of images visually similar to the input image.
- 37
- Predicting memorability of images using attention-driven spatial pooling and image semantics34
This article presents an attention-driven spatial pooling strategy for image memorability and shows that image features which are automatically extracted from images provide memorability predictions as nearly accurate as those derived from human annotations.
- Top down saliency estimation via superpixel-based discriminative dictionaries33
This work presents a novel method for learning top-down visual saliency, which is well-suited to locate objects of interest in complex scenes and provides much better saliency maps.
- EVREAL: Towards a Comprehensive Benchmark and Analysis Suite for Event-based Video Reconstruction32
This paper proposes a unified evaluation methodology and introduces an open-source framework called EVREAL to comprehensively benchmark and analyze various event-based video reconstruction methods from the literature and provides valuable insights into the performance of these methods under varying settings, challenging scenarios, and downstream tasks.
- CLIP-guided StyleGAN Inversion for Text-driven Real Image Editing32
CLIPInverter is presented, a new text-driven image editing approach that is able to efficiently and reliably perform multi-attribute changes and outperforms competing approaches in terms of manipulation accuracy and photo-realism on various domains including human faces, cats, and birds.
- TasvirEt: A benchmark dataset for automatic Turkish description generation from images30
The findings indicate that the new Turkish dataset and the approaches used here can be successfully used for automatically describing images in Turkish.
- ViLMA: A Zero-Shot Benchmark for Linguistic and Temporal Grounding in Video-Language Models29
It is shown that current VidLMs' grounding abilities are no better than those of vision-language models which use static images, and this benchmark serves as a catalyst for future research on VidL Ms, helping to highlight areas that still need to be explored.
- Manipulating Attributes of Natural Scenes via Hallucination29
A two-stage framework for enabling users to directly manipulate high-level attributes of a natural scene using a deep generative network that can hallucinate images of a scene as if they were taken in a different season, weather condition, or at a different time of the day.
- 29
- Burst Photography for Learning to Enhance Extremely Dark Images28
This paper proposes a novel coarse-to-fine network architecture that generates high-quality outputs progressively and extends this network to a permutation invariant structure so that it takes a burst of low-light images as input and merges information from multiple images at the feature-level.
- 28
- Two-Stream Convolutional Networks for Dynamic Saliency Prediction.28
The experiments show that taking motion into account in this way can be helpful for static saliency estimation, and a novel, empirically grounded data augmenta- tion technique is introduced for this task.
- 26
- Automatic Description Generation from Images: A Survey of Models, Datasets, and Evaluation Measures (Extended Abstract)25
An overview of the benchmark image-text datasets and the evaluation measures that have been developed to assess the quality of machine-generated descriptions and future directions in the area of automatic image description are explored.
- SonicDiffusion: Audio-Driven Image Generation and Editing with Pretrained Diffusion Models22
This work proposes a method to enable audio-conditioning in large scale image diffusion models by first maps features obtained from audio clips to tokens that can be injected into the diffusion model in a fashion similar to text tokens.
- Object and relation centric representations for push effect prediction22
This paper proposes a graph neural network based framework for effect prediction and parameter estimation of pushing actions by modeling object relations based on contacts or articulations that enables the robot to predict and adapt the effect of a pushing action as it observes the scene.
- Hippocrates: An Open-Source Framework for Advancing Large Language Models in Healthcare21
Hippocrates is presented, an open-source LLM framework specifically developed for the medical domain, which offers unrestricted access to its training datasets, codebase, checkpoints, and evaluation protocols and introduces Hippo, a family of 7B models tailored for the medical domain.
- Leveraging semantic saliency maps for query-specific video summarization21
This work presents a weakly supervised query-focused video summarization method that makes use of semantic attributes as an indicator of query relevance and semantic attention maps to locate related regions in the frames and utilizes both within a submodular maximization framework.
- Multiple-Instance Learning with Instance Selection via Dominant Sets19
This paper addresses MIL from a view that transforms the problem into a standard supervised learning problem via instance selection, based on an effective pairwise clustering algorithm referred to as dominant sets.
- A similarity-based approach for shape classification using Aslan skeletons19
This study presents a similarity-based approach for classifying 2D shapes based on their Aslan skeletons, and the coarse structure of this skeleton representation allows us to represent each shape category in the form of a reduced set of prototypical trees.
- Using synthetic data for person tracking under adverse weather conditions18
The experimental results demonstrate that the performances of the state-of-the-art deep trackers under adverse weather conditions can be boosted when the available real training sequences are complemented with the authors' synthetically generated dataset during training.
- ST360IQ: No-Reference Omnidirectional Image Quality Assessment With Spherical Vision Transformers17
This study presents a method for no-reference 360° image quality assessment that employs a vision-transformers based module processing saliency selective patches/tokens that estimates a quality score from each viewport and aggregates these scores to give a final quality score.
- Belief Regulated Dual Propagation Nets for Learning Action Effects on Groups of Articulated Objects17
BRDPN is introduced, a general-purpose learnable physics engine, which enables a robot to predict the effects of its actions in scenes containing groups of articulated multi-part objects and can adapt its physic predictions, since the relations can be predicted online.
- A Gated Fusion Network for Dynamic Saliency Prediction15
This article introduces the GFSalNet, the first deep saliency model capable of making predictions in a dynamic way via the gated fusion mechanism, and shows that it has a good generalization ability, and exploits temporal information more effectively via its adaptive fusion scheme.
- GaussianVideo: Efficient Video Representation via Hierarchical Gaussian Splatting14
A novel neural video representation that combines 3D Gaussian splatting with continuous camera motion modeling that achieves high-quality rendering at impressive speeds by leveraging Neural ODEs and introducing a spatiotemporal hierarchical learning strategy.
- Leveraging auxiliary image descriptions for dense video captioning14
This paper uses a generator and discriminator design, together with an attention-based fusion technique, to incorporate image captions as context in the video caption generation process and achieves more accurate and more diverse video descriptions.
- 13
- Procedural Reasoning Networks for Understanding Multimodal Procedures13
This paper introduces a new entity-aware neural comprehension model augmented with external relational memory units and finds that this model learns effective dynamic representations of entities even though it does not use any supervision at the level of entity states.
- 12
- Data‐driven image captioning via salient region discovery12
An object-based semantic image representation is integrated into a deep features-based retrieval framework to select the relevant images and a novel phrase selection paradigm and sentence generation model which depends on a joint analysis of salient regions in the input and retrieved images within a clustering framework are presented.
- A comparative study for feature integration strategies in dynamic saliency estimation12
This paper investigates four learning-based (SVM, Gradient Boosting, NNLS, Random Forest) and two transformation- based (Mean, Max) fusion methods, resulting in six new dynamic saliency models, which achieve better performance than the individual features.
- NOVA: Rendering Virtual Worlds with Humans for Computer Vision Tasks11
NOVA, a versatile framework to create realistic-looking 3D rendered worlds containing procedurally generated humans with rich pixel-level ground truth annotations, is presented, indicating that the synthetic data generated by NOVA represents a good proxy of the real-world and can be exploited for computer vision tasks.
- MSVD-Turkish: a comprehensive multimodal video dataset for integrated vision and language research in Turkish10
This work creates the first large-scale video captioning dataset for Turkish by carefully translating the English descriptions of the videos in the MSVD (Microsoft Research Video Description Corpus) dataset into Turkish.
- 10
- Computer vision based unistroke keyboard system and mouse for the handicapped10
A disabled person who has difficulty using his/her hands may attach the laser pointer to an eyeglass and easily enter text by moving his/Her head to point the laser beam on a character location.
- Joint Exploitation of Features and Optical Flow for Real-Time Moving Object Detection on Drones9
This paper introduces sparse flow based parallax handling and downscale processing and pushes the boundaries of real-time performance with 16 FPS on limited embedded resources (a five-fold improvement over existing baselines), while managing to perform comparably or even improve the state-of-the-art in two different datasets.
- Spherical Vision Transformer for 360-degree Video Saliency Prediction8
This work proposes a novel vision-transformer-based model for omnidirectional videos named SalViT360 that leverages tangent image representations and introduces a spherical geometry-aware spatiotemporal self-attention mechanism that is capable of effective omnid Directional video understanding.
- Omnidirectional image quality assessment with local–global vision transformers7
The Local-Global Transformer for 360-degree Image Quality Assessment (LGT360IQ) is introduced, a novel framework that combines local and global information, offering an effective IQA method for ODIs.
- 7
- HyperGAN-CLIP: A Unified Framework for Domain Adaptation, Image Synthesis and Manipulation6
This work presents a novel framework that significantly extends the capabilities of a pre-trained StyleGAN by integrating CLIP space via hypernetworks and introduces a CLIP-guided discriminator that enhances the alignment between generated images and target domains, ensuring superior image quality.
- Synthetic18K: Learning better representations for person re-ID and attribute recognition from 1.4 million synthetic images6
This paper introduces Synthetic18K, a large-scale dataset of over 1 million computer generated person images of 18K unique identities with relevant attributes and demonstrates that pretraining of simple deep architectures on Synthetic 18K for person re-identification and attribute recognition and then fine-tuning on real data leads to significant improvements in prediction performances.
- 6
- 6
- 5
- 5
- Generating visual story graphs with application to photo album summarization5
The proposed approach identifies coherent visual storylines and exploits sub-modularity to select a subset of these lines which covers the general narrative at most and delivers better performance on next image prediction and coverage tasks than the state-of-the-art.
- HUCVL at MediaEval 2016: Predicting Interesting Key Frames with Deep Models.5
This work proposes three different deep models based on fine-tuning two pretrained models, namely AlexNet and MemNet, which depend on a triplet network which is comprised of three instances of the same feedforward network with shared weights, and trained according to a doublet ranking loss.
- Evaluating deghosting algorithms for HDR images5
Two recent approaches addressing this issue are compared using a novel dataset comprised of image sequences with varying complexity and the results are evaluated by using both objective and subjective measures.
- Edge Strength Functions as Shape Priors in Image Segmentation5
This work presents a new image segmentation method which accounts for prior shape information, which depends on Ambrosio-Tortorelli approximation of Mumford-Shah functional and can handle arbitrary deformations due to shape variability as well as plane Euclidean transformations.
- LAMP: Language-Assisted Motion Planning for Controllable Video Generation4
This work introduces LAMP, the first framework for generating both object and camera motions directly from natural language specifications, and constructs a large-scale procedural dataset pairing natural text descriptions with corresponding motion programs and 3D trajectories.
- 4
- From Noon to Sunset: Interactive Rendering, Relighting, and Recolouring of Landscape Photographs by Modifying Solar Position4
This paper leverages a precomputed atmospheric scattering algorithm to make all of the changes on the position of the sun not only realistic but also in real‐time.
- A Game Theoretic Approach to Learning Shape Categories and Contextual Similarities4
Experiments showed that the jointly learning the categories from examples and the similarity measures related to them approach outperform other clustering approaches that do not make use of the underlying contextual information and provides similarities comparable with a state-of-the-art label-propagation approach which, however, cannot extract categories.
- VidStyleODE: Disentangled Video Editing via StyleGAN and NeuralODEs3
VidStyleODE encodes the video content in a pre-trained StyleGAN space and benefits from a latent ODE component to summarize the spatiotemporal dynamics of the input video, and combines the two to generate high-quality and temporally consistent videos with varying frame rates.
- Perception-Distortion Trade-Off in the SR Space Spanned by Flow Models3
This paper proposes different image ensembling and fusion strategies which offer multiple paths to move sample solutions in the SR space to more desired destinations in the perception-distortion plane in a controllable manner depending on the fidelity vs. perceptual quality requirements of the task at hand.
- 3
- Vision-based continuous GraffitiTM-like text entry system3
A vision-based continuous Graffiti™-like text entry system is presented that sketches characters in a Graffiti-like alphabet in a continuous manner on a flat surface using a laser pointer.
- Beyond Gaussian Bottlenecks: Topologically Aligned Encoding of Vision-Transformer Feature Spaces2
This work introduces a novel type of variational autoencoder using a product of Power Spherical latent distributions, explicitly enforcing hyperspherical structure in the bottleneck to preserve directional and geometric semantics under strong compression.
- 2
- Spherical Vision Transformers for Audio-Visual Saliency Prediction in 360-Degree Videos2
The results on a number of benchmark datasets demonstrate that SalViT360 and SalViT360-AV significantly outperform existing methods in predicting viewer attention in 360-degree scenes, and suggest that integrating spatial audio cues in the model architecture is crucial for accurate saliency prediction in omnidirectional videos.
- Evaluating Linguistic Capabilities of Multimodal LLMs in the Lens of Few-Shot Learning2
This study aims to evaluate the performance of MLLMs on the VALSE benchmark, focusing on the efficacy of few-shot In-Context Learning (ICL), and Chain-of-Thought (CoT) prompting.
- 2
- Modulating Bottom-Up and Top-Down Visual Processing via Language-Conditional Filters2
This work argues that the common practice of using language in a top-down manner, to direct visual attention over high-level visual features, may not be optimal and hypothesizes that the use of language to also condition the bottom-up processing from pixels to high- level features can provide benefits to the overall performance.
- Belief Regulated Dual Propagation Nets for Learning Action Effects on Articulated Multi-Part Objects.2
This paper introduces Belief Regulated Dual Propagation Networks (BRDPN), a general purpose learnable physics engine, which enables a robot to predict the effects of its actions in scenes containing groups of articulated multi-part objects.
- 2
- Leveraging Captions in the Wild to Improve Object Detection2
It is shown how a pre-trained state-of-the-art object detector can take advantage of noisy captions and utilize them to improve object detection.
- 2
- Data-driven image captioning with meta-class based retrieval2
This study presents a novel data-driven image captioning strategy which, for a given image, finds the most visually similar image in a large dataset of image-caption pairs and transfers its caption as the description of the input image.
- A dynamic procedure for forming shape categories from skeletal shape trees2
This report presents a dynamic procedure for the formation of category trees from a collection of skeletal shape trees that overcomes several shortcomings of the formation procedure proposed by Baseski [4].
- Auteur: Language-Driven Cinematographic Framing for Human-Centric Video Generation1
Auteur enables cinematographic framing of human-centered scenes, a capability largely absent in prior generative models, and is presented as a method for language-driven, human-centric camera framing in generative video.
- DeVisE: Behavioral Testing of Medical Large Language Models1
DeVisE ( De mographics and Vi tal s igns E valuation), a behavioral testing framework for probing fine-grained clinical understanding, is introduced, highlighting the utility of behavioral testing in exposing the reasoning strategies of clinical LLMs and informing the design of safer, more transparent medical AI systems.
- A Vision-Language Framework for Multispectral Scene Representation Using Language-Grounded Features1
This work presents a vision-language framework named Spectral-LLaVA, which integrates multispectral data with vision-language alignment techniques to enhance scene representation and description and highlights Spectral-LLaVA’s ability to produce detailed and accurate descriptions.
- 1
- 1
- 1
- Disentangling Content and Motion for Text-Based Neural Video Manipulation1
This paper introduces a new method called DiCoMoGAN for manipulating videos with natural language, aiming to perform local and semantic edits on a video clip to alter the appearances of an object of interest.
- 1
- 1
- 1
- 1
- 1
- Coarse-to-Fine Matching of Shapes Using Disconnected Skeletons by Learning Class-Specific Boundary Deformations1
This study presents a simple way to enrich disconnected skeletons with radius functions and attempts to resolve the conflicting goals of stability and sensitivity by proposing a coarse-to-fine shape matching algorithm.
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Hedging static saliency models to predict dynamic saliency–
A novel weakly supervised dynamic saliency model called HedgeSal is proposed, which is based on a decision-theoretic online learning scheme and demonstrates the effectiveness of the model on the CRCNS, UCFSports and CITIUS datasets.
- –
- –
- –
- Structure-Texture Decomposition of RGB-D Images–
This paper proposes to incorporate three geometric depth features, namely height above ground, angle with gravity and horizontal disparity to the pool of image features used in that study, and proposes to use a new kernel function based on KL-divergence between the distributions of extracted features.
- –
- –
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.