Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks235 from public data
- Dynamic Convolution: Attention Over Convolution Kernels1,508
By simply using dynamic convolution for the state-of-the-art architecture MobileNetV3-Small, the top-1 accuracy of ImageNet classification is boosted by 2.9% with only 4% additional FLOPs and 2.8 AP gain is achieved on COCO keypoint detection.
- CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windows1,458
The Cross-Shaped Window self-attention mechanism for computing self-Attention in the horizontal and vertical stripes in parallel that form a cross-shaped window is developed, with each stripe obtained by splitting the input feature into stripes of equal width.
- Vector Quantized Diffusion Model for Text-to-Image Synthesis1,117
This method is based on a vector quantized variational autoencoder whose latent space is modeled by a conditional variant of the recently developed Denoising Diffusion Probabilistic Model (DDPM), and it is found that this latent-space method is well-suited for text-to-image generation tasks.
- Dynamic Head: Unifying Object Detection Heads with Attentions1,037
This paper presents a novel dynamic head framework to unify object detection heads with attentions by coherently combining multiple self-attention mechanisms between feature levels for scale- awareness, among spatial locations for spatial-awareness, and within output channels for task-awareness that significantly improves the representation ability of object detection Heads without any computational overhead.
- Multi-attentional Deepfake Detection995
A new multi-attentional deepfake detection network that consists of three key components: multiple spatial attention heads to make the network attend to different local parts, a new regional independence loss and an attention guided data augmentation strategy, and state-of-the-art performance.
- Gated Context Aggregation Network for Image Dehazing and Deraining803
An end-to-end gated context aggregation network to directly restore the final haze-free image by adopting the latest smoothed dilation technique to help remove the gridding artifacts caused by the widely-used dilated convolution with negligible extra parameters.
- Mobile-Former: Bridging MobileNet and Transformer727
The proposed light-weight cross attention to model the bridge enables bidirectional fusion of local and global features and builds an efficient end-to-end detector by replacing backbone, encoder and decoder in DETR with Mobile-Former, which outperforms DETR but saves 52% of computational cost and 36% of parameters.
- StyleBank: An Explicit Representation for Neural Image Style Transfer504
This work proposes StyleBank, which is composed of multiple convolution filter banks and each filter bank explicitly represents one style, for neural image style transfer, the first style transfer network that links back to traditional texton mapping methods, and hence provides new understanding on neural style transfer.
- CLIP-NeRF: Text-and-Image Driven Manipulation of Neural Radiance Fields490
This work introduces a disentangled conditional NeRF architecture that allows individual control over both shape and appearance and proposes an inverse optimization method that accurately projects an input image to the latent codes for manipulation to enable editing on real images.
- Deep exemplar-based colorization342
This work proposes the first deep learning approach for exemplar-based local colorization, which performs robustly and generalizes well even when using reference images that are unrelated to the input grayscale image.
- 326
- High-Fidelity Pluralistic Image Completion with Transformers314
The proposed method vastly outperforms state-of-the-art methods in terms of large performance boost on image fidelity even compared to deterministic completion methods; better diversity and higher fidelity for pluralistic completion; and exceptional generalization ability on large masks and generic dataset, like ImageNet.
- Coherent Online Video Style Transfer313
This work proposes the first end-toend network for online video style transfer, which generates temporally coherent stylized video sequences in near realtime and can achieve visually comparable coherence to optimization-based video styleTransfer, but is three orders of magnitude faster.
- Equivariant Multi-Modality Image Fusion299
This work proposes the Equivariant Multi-Modality imAge fusion (EMMA) paradigm, a novel training paradigm that encompasses a fusion module, a pseudo-sensing module, and an equivariant fusion module that enables the net training to follow the principles of the natural sensing-imaging process while satisfying the equivariant imaging prior.
- MaskCLIP: Masked Self-Distillation Advances Contrastive Language-Image Pretraining294
This paper presents a simple yet effective framework MaskCLIP, which incorporates a newly proposed masked self-distillation into contrastive language-image pretraining and introduces the local semantic supervision into the text branch, which further improves the pretraining performance.
- PeCo: Perceptual Codebook for BERT Pre-training of Vision Transformers289
A self-supervised transformer model for deep feature extraction and it is shown that it works well for calculating perceptual similarity, and such learned visual tokens indeed exhibit better semantic meanings, and help pre-training achieve superior transfer performance in various downstream tasks.
- General Facial Representation Learning in a Visual-Linguistic Manner281
This paper studies the transfer performance of pre-trained models on face analysis tasks and introduces a framework, called FaRL, for general facial representation learning, which achieves better transfer performance compared with previous pre- trained models and achieves its superiority in the low-data regime.
- BEVT: BERT Pretraining of Video Transformers279
BEVT is introduced which decouples video representation learning into spatial representation learning and temporal dynamics learning and achieves state-of-the-art performance on three challenging video benchmarks where BEVT achieves very promising results.
- Bringing Old Photos Back to Life261
A novel triplet domain translation network by leveraging real photos along with massive synthetic image pairs to restore old photos that suffer from severe degradation through a deep learning approach outperforms state-of-the-art methods in terms of visual quality.
- Dynamic ReLU226
This paper proposes dynamic ReLU (DY-ReLU), a dynamic rectifier of which parameters are generated by a hyper function over all in-put elements, and encodes the global context into the hyper function, and adapts the piecewise linear activation function accordingly.
- Unsupervised Pre-training for Person Re-identification214
This paper presents a large scale unlabeled person re-identification (Re-ID) dataset "LUPerson" and makes the first attempt of performing unsupervised pre-training for improving the generalization ability of the learned person Re-ID feature representation.
- Protecting Celebrities from DeepFake with Identity Consistency Transformer194
It is shown that Identity Consistency Transformer exhibits superior generalization ability not only across different datasets but also across various types of image degradation forms found in real-world applications including deepfake videos.
- NeRF-Art: Text-Driven Neural Radiance Fields Stylization183
This paper presents NeRF-Art, a text-guided NeRF stylization approach that manipulates the style of a pre-trained NeRF model with a simple text prompt that can shift a 3D scene to the target style characterized by desired geometry and appearance variations without any mesh guidance.
- Deep Model Intellectual Property Protection via Deep Watermarking176
A new model watermarking framework is proposed for protecting deep networks trained for low-level computer vision or image processing tasks, which can resist attacks with different network structures and objective functions.
- Model Watermarking for Image Processing Networks170
The first model watermarking framework for protecting image processing models, which can resist surrogate models learned with different network structures and objective functions is proposed and is easy to be extended to protect data and traditional image processing algorithms.
- A General Decoupled Learning Framework for Parameterized Image Operators161
This work proposes a new decoupled learning algorithm to learn from the operator parameters to dynamically adjust the weights of a deep network for image operators, denoted as the base network, and forms a network formed as another network, namely the weight learning network, which can be end-to-end jointly trained with the basenetwork.
- HairCLIP: Design Your Hair by Text and Reference Image147
This paper proposes a new hair editing interaction mode, which enables manipulating hair attributes individually or jointly based on the texts or reference images provided by users, and encode the image and text conditions in a shared embedding space and proposes a unified hair editing framework.
- Poison Ink: Robust and Invisible Backdoor Attack139
This work proposes a robust and invisible backdoor attack called Poison Ink, which is not only general to different datasets and network architectures but also flexible for different attack scenarios and has very strong resistance against many state-of-the-art defense techniques.
- Masked Video Distillation: Rethinking Masked Feature Modeling for Self-supervised Video Representation Learning135
This paper proposes masked video distillation (MVD), a simple yet effective two-stage masked feature modeling framework for video representation learning that demonstrates that video transformers pre-trained with spatial-temporal co-teaching outperform models distilled with a single teacher on a multitude of video datasets.
- LG-GAN: Label Guided Adversarial Network for Flexible Targeted Attack of Point Cloud Based Deep Networks130
The proposed LG-GAN can support flexible targeted attack on the fly while guaranteeing good attack performance and higher efficiency simultaneously, and is the first generation based 3D point cloud attack method.
- Shape-invariant 3D Adversarial Point Clouds120
A novel Point-Cloud Sensitivity Map is proposed to boost both the efficiency and imperceptibility of point perturbations and reveals the vulnerability of point cloud recognition models when encountering shape-invariant adversarial noises.
- 120
- 117
- Stereoscopic Neural Style Transfer117
This paper presents the first attempt at stereoscopic neural style transfer, which responds to the emerging demand for 3D movies or AR/VR, and proposes the first feed-forward network by jointly training a stylization sub- network and a disparity sub-network, and integrate them in a feature level middle domain.
- Efficient Semantic Image Synthesis via Class-Adaptive Normalization114
A truly spatially-adaptive variant of CLADE, namely CLADE-ICPE, which can be generalized to different SPADE-based methods while achieving comparable generation quality compared to SPADE, but it is much more efficient with fewer extra parameters and lower computational cost.
- MicroNet: Improving Image Recognition with Extremely Low FLOPs113
This paper found that two factors, sparse connectivity and dynamic activation function, are effective to improve the accuracy and proposes micro-factorized convolution, which factorizes a convolution matrix into low rank matrices, to integrate sparse connectivity into convolution.
- Online multi-object tracking with unsupervised re-identification learning and occlusion estimation110
This study shows that, when applied to state-of-the-art MOT methods, the proposed unsupervised re-Identification learning is comparable to supervised re-identification learning, and the tracking performance is further improved by the proposed occlusion estimation module.
- AvatarCraft: Transforming Text into Neural Human Avatars with Parameterized Shape and Pose Control108
The proposed method, AvatarCraft, addresses the challenge of creating a 3D human avatar with a specific identity and artistic style that can be easily animated by using diffusion models to guide the learning of geometry and texture for a neural avatar based on a single text prompt.
- Meta-PU: An Arbitrary-Scale Upsampling Network for Point Cloud106
A novel method called “Meta-PU” is proposed to first support point cloud upsampling of arbitrary scale factors with a single model and even outperforms the existing methods trained for a specific scale factor only.
- Reduce Information Loss in Transformers for Pluralistic Image Inpainting104
A patch-based auto-encoder P-VQVAE, where the encoder converts the masked image into non-overlapped patch tokens and the decoder recovers the masked regions from the inpainted tokens while keeping the unmasked regions unchanged is designed.
- Progressive Color Transfer With Dense Semantic Correspondences103
This work proposes a new algorithm for color transfer between images that have perceptually similar semantic structures that uses neural representations for matching and optimizes matching and color transfer, adopting a coarse-to-fine strategy.
- Diversity-Aware Meta Visual Prompting99
DAM-VP is an efficient and effective prompting method for transferring pre-trained models to downstream tasks with frozen backbone, clearly surpassing previous prompting methods in a series of downstream datasets for different pretraining models.
- Optochemogenetic Stimulation of Transplanted iPS-NPCs Enhances Neuronal Repair and Functional Recovery after Ischemic Stroke93
This study pioneers the generation and application of an optochemogenetics approach in stem cell transplantation therapy after stroke for optimal neural repair and functional recovery and shows enhanced regenerative properties of LMO3-iPS-NPCs in vitro and after transplantation into the ischemic brain of different genders and ages.
- 83
- Large-Scale Pre-training for Person Re-identification with Noisy Labels82
This paper develops a large-scale Pre-training frame-work utilizing Noisy Labels (PNL), which consists of three learning modules: supervised Re-ID learning, prototype-based contrastive learning, and label-guided Contrastive learning.
- Diverse Semantic Image Synthesis via Probability Distribution Modeling81
This paper proposes a novel diverse semantic image synthesis framework from the perspective of semantic class distributions, which naturally supports diverse generation at semantic or even instance level by modeling class-level conditional modulation parameters as continuous probability distributions instead of discrete values.
- SinDiffusion: Learning a Diffusion Model From a Single Natural Image80
It is uncovered that multiple models at progressive growing scales are not essential for learning from a single image and proposed SinDiffusion, a single diffusion-based model trained on a single scale, which is better-suited for this task.
- 79
- MichiGAN73
MichiGAN (Multi-Input-Conditioned Hair Image GAN), a novel conditional image generation method for interactive portrait hair manipulation that allows fully-conditioned hair generation from multiple user inputs, is presented.
- 72
- Simultaneous determination of 16 synthetic colorants in hotpot condiment by high performance liquid chromatography72
This method is sensitive and reliable, and can be used to simultaneously determine 8 lipid- soluble and 8 water-soluble colorants in hotpot condiment.
- 66
- Bringing Old Films Back to Life66
A learning-based framework, recurrent transformer network (RTN), to restore heavily degraded old films, based on the hidden knowledge learned from adjacent frames that contain abundant information about the occlusion, which is beneficial to restore challenging artifacts of each frame while ensuring temporal coherency.
- 66
- Learning with Noisy Labels for Robust Point Cloud Segmentation65
This work proposes a novel point-wise confidence selection to obtain reliable labels based on the historical predictions of each point, and presents a novel cluster-wise label correction to generate the best possible label taking the neighbor point correlations into consideration.
- Decouple Learning for Parameterized Image Operators65
A new decouple learning algorithm to learn from the operator parameters to dynamically adjust the weights of a deep network for image operators, denoted as the base network is proposed.
- 64
- Look Before You Match: Instance Understanding Matters in Video Object Segmentation61
It is argued that instance understanding matters in VOS, and integrating it with memory-based matching can enjoy the synergy, which is intuitively sensible from the definition of VOS task, i.e., identifying and segmenting object instances within the video.
- 60
- Combined transplantation of mesenchymal stem cells and endothelial progenitor cells for tissue engineering: a systematic review and meta-analysis60
Compared with combined transplantation (EPCs and EPCs) and transplantation of MSCs alone, cotransplantation significantly promoted angiogenesis and bone regeneration, as well as vessel revascularization and tissue repair in cerebrovascular disease and during bone regeneration.
- 57
- 55
- Deep Decomposition Learning for Inverse Imaging Problems54
Inspired by the geometry that data can be decomposed by two components from the null-space of the forward operator and the range space of its pseudo-inverse, this work explicitly reformulate the neural network layers as learning range-nullspace decomposition functions with reference to the layer inputs, instead of learning unreferenced functions.
- 52
- 51
- 44
- Improved Image Matting via Real-time User Clicks and Uncertainty Estimation42
An improved deep image matting framework which is trimap-free and only needs several user click interactions to eliminate the ambiguity is proposed, and a new uncertainty estimation module that can predict which parts need polishing and a following local refinement module is introduced.
- 40
- 37
- 36
- Transformer Based Pluralistic Image Completion With Reduced Information Loss35
A patch-based auto-encoder that directly takes features from the P-VQVAE encoder as input without any quantization and only regards the quantized tokens as prediction targets is designed, which greatly outperforms existing transformer based methods on image fidelity.
- 34
- Density-Aware Graph for Deep Semi-Supervised Visual Recognition34
This paper proposes a new Density-aware Neighborhood Aggregation module to learn more discriminative features by incorporating the neighborhood information in a density-aware manner and a novel D density-ascending Path based Label Propagation module to generate the pseudo-labels for unlabeled samples more efficiently according to the feature distribution characterized by density.
- Current Status and Analysis of Machine Learning in Hepatocellular Carcinoma33
The application of different ML models in HCC is explained, the overall prognosis remains poor, and suggestions for follow-up research are provided.
- Once a MAN: Towards Multi-Target Attack via Learning Multi-Target Adversarial Network Once33
This paper proposes the first Multi-target Adversarial Network (MAN), which can generate multi-target adversarial samples with a single model and can achieve better classification accuracy than other methods when attacked by various methods.
- Development of a simple method for simultaneous determination of nine subclasses of non-steroidal anti-inflammatory drugs in milk and dairy products by ultra-performance liquid chromatography with tandem mass spectrometry33
The results proved that this method was suitable for the simultaneous determination of nine subclasses of NSAIDs residues in milk and dairy products.
- HairCLIPv2: Unifying Hair Editing via Proxy Feature Blending31
Besides the unprecedented user interaction mode support, quantitative and qualitative experiments demonstrate the superiority of HairCLIPv2 in terms of editing effects, irrelevant attribute preservation and visual naturalness.
- 30
- 30
- 30
- 28
- 28
- 28
- 27
- FreeFlux: Understanding and Exploiting Layer-Specific Roles in RoPE-Based MMDiT for Versatile Image Editing24
This work presents the first mechanistic analysis of RoPE-based MMDiT models, introducing an automated probing strategy that disentangles positional information versus content dependencies by strategically manipulating RoPE during generation.
- 24
- 22
- 21
- 21
- 21
- 20
- 20
- 19
- 19
- 18
- 18
- 18
- 18
- 17
- 17
- 17
- 17
- Translation of Aerial Image Into Digital Map via Discriminative Segmentation and Creative Generation17
- 16
- 16
- 16
- Rethinking Visual Prompting for Multimodal Large Language Models with External Knowledge15
This paper proposes a new visual prompt approach to integrate fine-grained external knowledge, gleaned from specialized vision models, into MLLMs, and can be effortlessly incorporated into various MLLMs, such as LLaVA and Mipha, considerably improving their visual understanding performance.
- 15
- 15
- 14
- 14
- 14
- 13
- 13
- 13
- 12
- 12
- 12
- 12
- 12
- 12
- 12
- 12
- 11
- 11
- 11
- 11
- Screening and optimization of AFB1 aptamer for target-capture electrochemical aptasensor development10
- 10
- 10
- 10
- 10
- 10
- Adversarial defense via self-orthogonal randomization super-network10
This paper proposes a novel type of defense method named ‘‘Self-Orthogonal Randomization Super-network” ( SORS), which can achieve significantly better robustness than the vanilla single model baseline and the naive model ensemble baseline.
- 10
- 10
- 9
- 9
- 9
- 8
- 8
- Audio-Visual Contrastive Pre-train for Face Forgery Detection8
This work proposes a video-level deepfake detection method based on a temporal transformer with a self-supervised audio–visual contrastive learning approach for pre-training the deepfake detector that achieves better generalization ability in cross-dataset evaluations.
- 8
- 8
- 8
- 8
- 8
- 8
- 7
- 7
- 7
- 6
- 6
- 6
- 6
- 5
- Olympus: A Universal Task Router for Computer Vision Tasks5
Olympus is introduced, a new approach that transforms Multimodal Large Language Models (MLLMs) into a unified framework capable of handling a wide array of computer vision tasks, and easily integrates with existing MLLMs, expanding their capabilities with comparable performance.
- 5
- 5
- 5
- 5
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- 4
- Pluralistic Salient Object Detection3
Two new SOD datasets “DUTS-MM” and “DUTS-MQ” are presented, along with newly designed evaluation metrics, and a simple yet effective pluralistic SOD baseline based on a Mixture-of-Experts (MOE) design is proposed.
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- : Color decomposition-based adversarial examples for screen devices2
Inspired by the properties of the human visual system, the first color decomposition-based adversarial example method CDAE for screen devices is designed, which decomposes one regular screen frame into two symmetric adversarial frames with maximum modifications while theoretically guaranteeing the visual quality perceived by human observers.
- 2
- 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.