Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks107 from public data
- Q-Align: Teaching LMMs for Visual Scoring via Discrete Text-Defined Levels664
The proposed Q-Align achieves state-of-the-art performance on image quality assessment (IQA), image aesthetic assessment (IAA), as well as video quality assessment (VQA) tasks under the original LMM structure and unify the three tasks into one model, termed the OneAlign.
- Cascade Residual Learning: A Two-Stage Convolutional Neural Network for Stereo Matching502
A novel cascade CNN architecture composing of two stages that advances the recently proposed DispNet by equipping it with extra up-convolution modules, leading to disparity images with more details, and shows that residual learning provides more effective refinement.
- GRNet: Gridding Residual Network for Dense Point Cloud Completion468
This work devise two novel differentiable layers, named Gridding and Gridding Reverse, to convert between point clouds and 3D grids without losing structural information, and presents the differentiable Cubic Feature Sampling layer to extract features of neighboring points, which preserves context information.
- Exploring Video Quality Assessment on User Generated Contents from Aesthetic and Technical Perspectives463
The Disentangled Objective Video Quality Evaluator (DOVER) is proposed, the first approach to provide reliable clear-cut quality evaluations from a single aesthetic or technical perspective, and the first approach to provide reliable clear-cut quality evaluations from a single aesthetic or technical perspective.
- TOPIQ: A Top-Down Approach From Semantics to Distortions for Image Quality Assessment423
This work proposes a top-down approach that uses high-level semantics to guide the IQA network to focus on semantically important local distortion regions, named as TOPIQ, and designs a heuristic coarse-to-fine network that leverages multi-scale features and progressively propagates multi-level semantic information to low-level representations in a top-down manner.
- Q-Bench: A Benchmark for General-Purpose Foundation Models on Low-level Vision309
Q-Bench, a holistic benchmark crafted to systematically evaluate potential abilities of MLLMs on three realms: low-level visual perception, low-level visual description, and overall visual quality assessment, confirms that MLLMs possess preliminary low-level visual skills.
- Accurate Single Stage Detector Using Recurrent Rolling Convolution297
A novel single stage end-to-end trainable object detection network is proposed by introducing Recurrent Rolling Convolution (RRC) architecture over multi-scale feature maps to construct object classifiers and bounding box regressors which are deep in context.
- 239
- Q-Instruct: Improving Low-Level Visual Abilities for Multi-Modality Foundation Models218
The first dataset consisting of human natural language feedback on low-level vision, and a GPT-participated transformation to convert these feedbacks into a rich set of 200K instruction-response pairs, termed Q-Instruct are collected.
- Single View Stereo Matching209
This paper shows for the first time that the monocular depth estimation problem can be reformulated as two sub-problems, a view synthesis procedure followed by stereo matching, with two intriguing properties, namely i) geometrical constraints can be explicitly imposed during inference; ii) demand on labelled depth data can be greatly alleviated.
- Shepard convolutional neural networks190
This paper draws on Shepard interpolation and design Shepard Convolutional Neural Networks (ShCNN) which efficiently realizes end-to-end trainable TVI operators in the network and shows that by adding only a few feature maps in the new Shepard layers, the network is able to achieve stronger results than a much deeper architecture.
- Efficient Regional Memory Network for Video Object Segmentation183
The proposed RM-Net effectively alleviates the ambiguity of similar objects in both memory and query frames, which allows the information to be passed from the regional memory to the query region efficiently and effectively.
- FuseFormer: Fusing Fine-Grained Information in Transformers for Video Inpainting180
This work proposes FuseFormer, a Transformer model designed for video inpainting via fine-grained feature fusion based on novel Soft Split and Soft Composition operations, which surpasses state-of-the-art methods in both quantitative and qualitative evaluations.
- Polarized Reflection Removal With Perfect Alignment in the Wild152
This work identifies the misalignment issues of existing reflection removal datasets where the collected reflection-free images are not perfectly aligned with input mixed images due to glass refraction and proposes a polarized reflection removal model with a two-stage architecture.
- LSTM Pose Machines148
It is shown that if the weight sharing scheme is imposed to the multi-stage CNN, it could be re-written as a Recurrent Neural Network (RNN), which decouples the relationship among multiple network stages and results in significantly faster speed in invoking the network for videos.
- Deep Animation Video Interpolation in the Wild127
This work formally defines and study the animation video interpolation problem for the first time, and proposes an effective framework, AnimeInterp, with two dedicated modules in a coarse-to-fine manner, which outperforms existing state-of-the-art interpolation methods for animation videos.
- Monocular Depth Estimation with Affinity, Vertical Pooling, and Label Enhancement124
This work explicitly model the relationships of different image locations with an affinity layer and combine absolute and relative features in an end-to-end network to obtain global understanding of the image depth and exploit residual learning to help depth refinement.
- Towards Real Scene Super-Resolution With Raw Images122
A new pipeline to generate realistic training data by simulating the imaging process of digital cameras is proposed and a dual convolutional neural network is developed to exploit the originally captured radiance information in raw images to recover fine details and clear structures.
- Look, Listen and Learn — A Multimodal LSTM for Speaker Identification115
A novel multimodal Long Short-Term Memory (LSTM) architecture which seamlessly unifies both visual and auditory modalities from the beginning of each sequence input is described which outperforms the state-of-the-art systems in speaker identification with lower false alarm rate and higher recognition accuracy.
- Enhanced Quadratic Video Interpolation100
This work improves the performance of QVI from three facets and proposes an enhanced quadratic video interpolation (EQVI) model that won the first place in the AIM2020 Video Temporal Super-Resolution Challenge and devise a novel multi-scale fusion network (MS-Fusion) which can be regarded as a learnable augmentation process.
- Neighbourhood Representative Sampling for Efficient End-to-End Video Quality Assessment99
This work proposes a unified scheme, spatial-temporal grid mini-cube sampling (St-GMS), and the resultant samples are named fragments, a network architecture tailored specifically for fragments that achieves up to 1612× efficiency than the existing state-of-the-art VQA benchmarks.
- DisCoVQA: Temporal Distortion-Content Transformers for Video Quality Assessment86
The proposed Temporal Distortion-Content Transformers for Video Quality Assessment (DisCoVQA) reaches state-of-the-art performance on several VQA benchmarks without any extra pre-training datasets and up to 10% better generalization ability than existing methods.
- Towards Explainable In-the-Wild Video Quality Assessment: A Database and a Language-Prompted Approach80
The MaxVQA is proposed, a language-prompted VQA approach that modifies vision-language foundation model CLIP to better capture important quality issues as observed in the authors' analyses.
- Deep Surface Normal Estimation With Hierarchical RGB-D Fusion79
A hierarchical fusion network with adaptive feature re-weighting is proposed for surface normal estimation from a single RGB-D image, outperforming state-of-the-art normal estimation schemes.
- 73
- Dual-Camera Super-Resolution with Aligned Attention Modules68
This work proposes a self-supervised domain adaptation strategy for real-world images that generalizes the standard patch-based feature matching with spatial alignment operations and further explores the dual-camera super-resolution that is one promising application of RefSR.
- Zoom and Learn: Generalizing Deep Stereo Matching to Novel Domains66
This work proposes a self-adaptation approach for CNN training, utilizing both synthetic training data and stereo pairs in the new domain (without ground-truths), and forms an iterative optimization problem with graph Laplacian regularization.
- StereoGAN: Bridging Synthetic-to-Real Domain Gap by Joint Optimization of Domain Translation and Stereo Matching52
The proposed end-to-end training framework introduces two novel losses, i.e., bidirectional multi-scale feature re-projection loss and correlation consistency loss, to help translate all synthetic stereo images into realistic ones as well as maintain epipolar constraints.
- Learning Spatial and Spatio-Temporal Pixel Aggregations for Image and Video Denoising51
A spatial pixel aggregation network is presented and the pixel sampling and averaging strategies for image denoising are learned and the proposed model naturally adapts to image structures and can effectively improve the denoised results.
- 46
- Deep Graph Laplacian Regularization for Robust Denoising of Real Images45
By integrating graph Laplacian regularization as a trainable module into a deep learning framework, this work is less susceptible to overfitting than pure CNN-based approaches, achieving higher robustness to small datasets and cross-domain denoising.
- 44
- 44
- An overview of free viewpoint depth-image-based rendering (DIBR)44
This paper discusses the techniques for depth map creation, the method for rendering given the depth information and post-processing for corresponding artifacts in the multi-view adaptation, and issues the quality measurement of the generated depth and synthesized view.
- Exploring the Effectiveness of Video Perceptual Representation in Blind Video Quality Assessment41
Experiments show that the perceptual representation in the HVS is an effective way of predicting subjective temporal quality, and thus TPQI can, for the first time, achieve comparable performance to the spatial quality metric and be even more effective in assessing videos with large temporal variations.
- 38
- Deep End-to-End Alignment and Refinement for Time-of-Flight RGB-D Module34
This work proposes a framework for jointly alignment and refinement via deep learning, achieving state-of-the-art for ToF refinement through improved kernel predicting network that performs kernel normalization and applies the bias prior to the dynamic convolution.
- 31
- Exploiting Raw Images for Real-Scene Super-Resolution31
This paper proposes a method to generate more realistic training data by mimicking the imaging process of digital cameras, and develops a two-branch convolutional neural network to exploit the radiance information originally-recorded in raw images.
- 31
- AIM 2020 Challenge on Video Temporal Super-Resolution29
This paper reports the second AIM challenge on Video Temporal Super-Resolution (VTSR), a.k.a. frame interpolation, with a focus on the proposed solutions, results, and analysis, and proposes the enhanced quadratic video interpolation method.
- SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecture28
This work introduces SenseNova-U1, a native unified multimodal paradigm built upon NEO-unify, in which understanding and generation evolve as synergistic views of a single underlying process that points toward a broader roadmap where models do not translate between modalities, but think and act across them in a native manner.
- 26
- 25
- Toward 3D object reconstruction from stereo images24
A new deep learning framework for reconstructing the 3D shape of an object from a pair of stereo images, which reasons about the3D structure of the object by taking bidirectional disparities and feature correspondences between the two views into account is proposed.
- AIM 2019 Challenge on Video Temporal Super-Resolution: Methods and Results22
This paper reviews the first AIM challenge on video temporal super-resolution (frame interpolation) with a focus on the proposed solutions and results, and employs the REDS_VTSR dataset derived from diverse videos captured in a hand-held camera for training and evaluation purposes.
- A Categorized Reflection Removal Dataset with Diverse Real-world Scenes21
It is shown that state-of-the-art reflection removal methods generally perform well on blurry reflection but fail in obtaining satisfying performance on other types of real-world reflection.
- Towards Robust Text-Prompted Semantic Criterion for In-the-Wild Video Quality Assessment19
An efficient fine-tuning scheme is devised for BVQI-Local that jointly optimizes text prompts and final fusion weights, resulting in state-of-the-art performance and superior generalization ability in comparison to prevalent opinion-driven VQA methods.
- 18
- MIPI 2023 Challenge on Nighttime Flare Removal: Methods and Results18
The developed solutions in this challenge achieved state-of-the-art performance on Nighttime Flare Removal and a detailed description of all models developed is provided in this paper.
- Semi-synthesis: A fast way to produce effective datasets for stereo matching15
This paper proposes semi-synthetic, an effective and fast way to synthesize large amount of data with closeto-real-scene texture to minimize the gap between synthetic data and real data.
- DSR: Direct Self-Rectification for Uncalibrated Dual-Lens Cameras15
This paper proposes a cost-effective solution to perform stereo rectification for dual-lens cameras called direct self-rectification, short for DSR, which removes the need of individual offline calibration for every pair of dual- lens cameras.
- 14
- 13
- Convolutional Memory Blocks for Depth Data Representation Learning12
A novel memory network module, called Convolutional Memory Block (CMB), is proposed, which empowers CNNs with the memory mechanism on handling depth data and is employed to develop a concise framework for predicting articulated pose from still depth images.
- 12
- 12
- 10
- 10
- 9
- 9
- 9
- A novel Ray-space based view generation algorithm via Radon transform9
A novel Radon transform based ray-space interpolation algorithm (RTI) is proposed to generate virtual viewpoint images in the EPI that can generate good interpolation result even if apriori depth range is not known.
- 9
- Deep Unrolled Weighted Graph Laplacian Regularization for Depth Completion8
The deep unrolled Weighted Graph Laplacian Regularization (WGLR) for depth completion which enhances input fidelity and noise robustness by enforcing input constraints in the network design and is integrated into deep learning framework to develop high-performance yet interpretable network.
- MIPI 2023 Challenge on RGBW Remosaic: Methods and Results8
This paper summarizes and reviews the RGBW Joint Remosaic and Denoise track on MIPI 2023, and describes the top three models developed in this challenge.
- 8
- 7
- 7
- 7
- 7
- 7
- MIPI 2023 Challenge on RGB+ToF Depth Completion: Methods and Results6
The results of the competition are presented, the strengths and weaknesses of the top-performing methods are analyzed, and the implications of the findings for future research in RGB+sparse ToF depth completion are discussed.
- 6
- Unsupervised Stereo Matching with Occlusion-Aware Loss6
This paper proposes an unsupervised stereo matching approach with a novel occlusion-aware reconstruction loss to enforce the disparity smoothness and correctness, and shows that the deep neural network can be well trained without requiring any ground-truth disparity data.
- 6
- 6
- Towards Geometry Guided Neural Relighting with Flash Photography5
This work proposes a framework for image relighting from a single flash photograph with its corresponding depth map using deep learning, and is able to extrapolate realistic high-frequency effects under novel lighting via geometry guided image decomposition from the flashlight image, and predict the cast shadow map from the shadow-encoding transformed depth map.
- 4
- 4
- Deep Graph Laplacian Regularization4
By integrating graph Laplacian regularization as a trainable module into a deep learning framework, this work is less susceptible to overfitting than pure CNN-based approaches, achieving higher robustness to small dataset and cross-domain denoising.
- 4
- Enhancing HDR Imaging with Joint Denoising and Deblurring3
A novel and efficient attention-based multi-exposure HDR imaging method, which skillfully selects the optimal information from corrupted LDR inputs by the authors' customized cross-attention mechanism to generate HDR information is proposed.
- 3
- MIPI 2023 Challenge on RGBW Fusion: Methods and Results3
This paper summarizes and reviews the RGBW Joint Fusion and Denoise track on MIPI 2023 and describes the models developed in this challenge, including PSNR, SSIM, LPIPS, and KLD.
- 3
- Texture optimization for seamless view synthesis through energy minimization3
The problem is formulated as joint optimization of inter-view texture similarity and geometry preservation, which significantly differs from traditional view synthesis framework, and the image characteristics of virtual view are inherently inherited from the reference view without introducing any image prior or texture modeling technique.
- AIDABench: AI Data Analytics Benchmark2
AIDABench is introduced, a comprehensive benchmark for evaluating AI systems on complex data analytics tasks in an end-to-end manner and offers a principled reference for enterprise procurement, tool selection, and model optimization.
- 2
- 1
- 1
- Matching Low-Quality Photo to DSLR-Quality with Deep Convolutional Networks1
This work proposes an end-to-end learning approach that translates low-quality photos captured by cameras with limited capabilities into DSLR-quailty photos and improves the existing state-of-art in terms of structural similarity measure, peak signal- to-noise ratio and by visual appearance.
- 1
- –
- –
- –
- –
- –
- –
- –
- –
- Inferring Depth from a Pair of Images Captured Using Different Aperture Settings–
Two possible ways of inferring depth of the scene from such an image pair with the constraint that both pictures are focused on the closest point of thescene are investigated.
- –
- –
- –
- –
- –
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.