Shijian Lu lists you as their PhD student. Claim this profile to confirm it and keep the rest of your record right.
Claim this profileAcademic lineage
View as a treeAdvisors
Works36 from public data
- Diverse Image Inpainting with Bidirectional and Autoregressive Transformers182
BAT-Fill is proposed, an innovative image inpainting framework that introduces a novel bidirectional autoregressive transformer (BAT) for image inPainting that incorporates the masked language model like BERT, which enables bidirectionally modeling of contextual information of missing regions for better image completion.
- WaveFill: A Wavelet-based Generation Network for Image Inpainting146
WaveFill is presented, a wavelet-based inpainting network that decomposes images into multiple frequency bands and fills the missing regions in each frequency band separately and explicitly and effectively mitigate inter-frequency conflicts while completing images in spatial domain.
- Accelerating DETR Convergence via Semantic-Aligned Matching136
SAM-DETR is presented, a Semantic-Aligned-Matching DETR that greatly accelerates DETR's convergence without sacrificing its accuracy and is like a plug and play, which complements existing convergence solutions well yet only introduces slight computational overhead.
- Multimodal Image Synthesis and Editing: The Generative AI Era132
This survey comprehensively contextualize the advance of the recent multimodal image synthesis and editing and formulate taxonomies according to data modalities and model types and provides insights about the current research challenges and possible directions for future research.
- StyleRF: Zero-Shot 3D Style Transfer of Neural Radiance Fields110
StyleRF (Style Radiance Fields), an innovative 3D style transfer technique that resolves the three-way dilemma by performing style transformation within the feature space of a radiance field, achieves superior 3D stylization quality with precise geometry reconstruction and it can generalize to various new styles in a zero-shot manner.
- EMLight: Lighting Estimation via Spherical Distribution Approximation77
A novel spherical mover's loss is designed that guides to regress light distribution parameters accurately by taking advantage of the subtleties of spherical distribution and under the guidance of the predicted spherical distribution, light intensity and ambient term, the neural projector synthesizes panoramic illumination maps with realistic light frequency.
- Unbalanced Feature Transport for Exemplar-based Image Translation74
A general image translation framework that incorporates optimal transport for feature alignment between conditional inputs and style exemplars in image translation and a semantic-activation normalization scheme that injects style features of exemplars into the image translation process successfully is presented.
- Modulated Contrast for Versatile Image Synthesis54
MoNCE is presented, a versatile metric that introduces image contrast to learn a calibrated metric for the perception of multifaceted inter-image distances and introduces optimal transport in MoNCE to modulate the pushing force of negative samples collaboratively across multiple contrastive objectives.
- Towards Counterfactual Image Manipulation via CLIP42
A novel contrastive loss is designed that exploits predefined CLIP-space directions to guide the editing toward desired directions from different perspectives and achieves accurate and realistic editing while driving by target texts with various counterfactual concepts.
- Marginal Contrastive Correspondence for Guided Image Generation40
A Marginal Contrastive Learning Network (MCL-Net) is designed that explores contrastive learning to learn domain-invariant features for realistic exemplar-based image translation and designs an innovative marginal contrastive loss that guides to establish dense correspondences explicitly.
- Bi-level Feature Alignment for Versatile Image Translation and Manipulation34
A versatile image translation and manipulation framework that achieves accurate semantic and style guidance in image generation by explicitly building a correspondence and designs a novel confidence feature injection module which mitigates mismatch problem by fusing features adaptively according to the reliability of built correspondences.
- KD-DLGAN: Data Limited Image Generation via Knowledge Distillation32
KD-DLGAN is proposed, a knowledge-distillation based generation framework that introduces pre-trained vision-language models for training effective data-limited generation models and achieves superior image generation with limited training data.
- WaveNeRF: Wavelet-based Generalizable Neural Radiance Fields31
WaveNeRF is designed, which integrates wavelet frequency decomposition into MVS and NeRF to achieve generalizable yet high-quality synthesis without any per-scene optimization, and achieves superior generalizable radiance field modeling when only given three images as input.
- 31
- VMRF: View Matching Neural Radiance Fields25
VMRF is designed, an innovative view matching NeRF that enables effective NeRF training without requiring prior knowledge in camera poses or camera pose distributions and outperforms the state-of-the-art qualitatively and quantitatively by large margins.
- Timeexpert: an Expert-Guided Video Llm for Video Temporal Grounding19
TimeExpert is introduced, a Mixture-of-Experts (MoE)-based Video-LLM that effectively decomposes VTG tasks by dynamically routing task-specific tokens to specialized experts, with increased computational efficiency.
- Pose-Free Neural Radiance Fields via Implicit Pose Regularization19
IR-NeRF is designed, an innovative pose-free NeRF that introduces implicit pose regularization to refine pose estimator with unposed real images and improve the robustness of the pose estimation for real images.
- 19
- Weakly Supervised 3D Open-vocabulary Segmentation18
This work distill open-vocabulary visual and textual knowledge from CLIP into a neural radiance field (NeRF) which effectively lifts 2D features into view-consistent 3D segmentation, suggesting that 3D open- Vocabulary segmentation can be effectively learned from 2D images and text-image pairs.
- CodeDance: A Dynamic Tool-integrated MLLM for Executable Visual Reasoning16
This work introduces CodeDance, which explores executable code as a general solver for visual reasoning, and empirically observe novel emergent behaviors during RL training, suggesting a general and scalable mechanism for executable visual reasoning.
- 15
- Debiasing Text-to-Image Diffusion Models13
This work formalizes the problem setting and uses the text descriptions of bias groups to establish an unsafe direction for guiding the diffusion process, and proposes an iterative distribution alignment (IDA) method, which shows efficiency and fast convergence in resolving the social bias in TTI diffusion models.
- POCE: Pose-Controllable Expression Editing13
POCE is presented, an innovative pose-controllable expression editing network that can generate realistic facial expressions and head poses simultaneously with just unpaired training images and can learn from unpaired face images effectively.
- ThinkGen: Generalized Thinking for Visual Generation12
ThinkGen is presented, the first think-driven visual generation framework that explicitly leverages MLLM's CoT reasoning in various generation scenarios, and proposes a separable GRPO-based training paradigm (SepGRPO), alternating reinforcement learning between the MLLM and DiT modules.
- Text4Seg++: Advancing Image Segmentation via Generative Language Modeling11
A novel text-as-mask paradigm that casts image segmentation as a text generation problem, eliminating the need for additional decoders and significantly simplifying the segmentation process is proposed, highlighting the effectiveness, scalability, and generalizability of text-driven image segmentation within the MLLM framework.
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.