Personalized text-to-image generation aims to synthesize novel images of a specific subject or style using only a few reference images. Recent methods based on Low-Rank Adaptation (LoRA) enable efficient single-concept customization by injecting lightweight, concept-specific adapters into pre-trained diffusion models. However, combining multiple LoRA modules for multi-concept generation often leads to identity missing and visual feature leakage. In this work, we identify two key issues behind these failures: (1) token-wise interference among different LoRA modules, and (2) spatial misalignment between the attention map of a rare token and its corresponding concept-specific region. To address these issues, we propose Token-Aware LoRA (TARA), which introduces a token mask to explicitly constrain each module to focus on its associated rare token to avoid interference, and a training objective that encourages the spatial attention of a rare token to align with its concept region. Our method enables training-free multi-concept composition by directly injecting multiple independently trained TARA modules at inference time. Experimental results demonstrate that TARA enables efficient multi-concept inference and effectively preserving the visual identity of each concept by avoiding mutual interference between LoRA modules.

Video outpainting is a challenging task that generates new video content by extending beyond the boundaries of an original input video, requiring both temporal and spatial consistency. Many existing methods utilize latent diffusion models with U-Net backbones but still struggle to achieve high quality and adaptability in generated content. Diffusion transformers (DiTs) have emerged as a promising alternative because of their superior performance. We introduce OutDreamer, a DiT-based video outpainting framework comprising two main components: a video control branch and a conditional outpainting branch. The video control branch effectively extracts masked video information, while the conditional outpainting branch generates missing content based on these extracted conditions. Additionally, we propose a mask-driven self-attention layer that dynamically integrates the given mask information, further enhancing the model's adaptability to outpainting tasks. Furthermore, we introduce a latent alignment loss to maintain overall consistency both within and between frames. For long video outpainting, we employ a cross-video-clip refiner to iteratively generate missing content, ensuring temporal consistency across video clips. Extensive evaluations demonstrate that our OutDreamer outperforms existing video outpainting methods on widely recognized benchmarks.

Vision-Language Models (VLMs) have demonstrated great potential in real-world applications. While existing research primarily focuses on improving their accuracy, the efficiency remains underexplored. Given the real-time demands of many applications and the high inference overhead of VLMs, efficiency robustness is a critical issue. However, previous studies evaluate efficiency robustness under unrealistic assumptions, requiring access to the model architecture and parameters-an impractical scenario in ML-as-a-service settings, where VLMs are deployed via inference APIs. To address this gap, we propose VLMInferSlow, a novel approach for evaluating VLM efficiency robustness in a realistic black-box setting. VLMInferSlow incorporates fine-grained efficiency modeling tailored to VLM inference and leverages zero-order optimization to search for adversarial examples. Experimental results show that VLMInferSlow generates adversarial images with imperceptible perturbations, increasing the computational cost by up to 128.47%. We hope this research raises the community's awareness about the efficiency robustness of VLMs.

Building on the success of diffusion models in image generation and editing, video editing has recently gained substantial attention. However, maintaining temporal consistency and motion alignment still remains challenging. To address these issues, this paper proposes DINO-guided Video Editing (DIVE), a framework designed to facilitate subject-driven editing in source videos conditioned on either target text prompts or reference images with specific identities. The core of DIVE lies in leveraging the powerful semantic features extracted from a pretrained DINOv2 model as implicit correspondences to guide the editing process. Specifically, to ensure temporal motion consistency, DIVE employs DINO features to align with the motion trajectory of the source video. For precise subject editing, DIVE incorporates the DINO features of reference images into a pretrained text-to-image model to learn Low-Rank Adaptations (LoRAs), effectively registering the target subject's identity. Extensive experiments on diverse real-world videos demonstrate that our framework can achieve high-quality editing results with robust motion consistency, highlighting the potential of DINO to contribute to video editing. Project page: https://dino-video-editing.github.io

Text-conditional image editing is a practical AIGC task that has recently emerged with great commercial and academic value. For real image editing, most diffusion model-based methods use DDIM Inversion as the first stage before editing. However, DDIM Inversion often results in reconstruction failure, leading to unsatisfactory performance for downstream editing. To address this problem, we first analyze why the reconstruction via DDIM Inversion fails. We then propose a new inversion and sampling method named Dual-Schedule Inversion. We also design a classifier to adaptively combine Dual-Schedule Inversion with different editing methods for user-friendly image editing. Our work can achieve superior reconstruction and editing performance with the following advantages: 1) It can reconstruct real images perfectly without fine-tuning, and its reversibility is guaranteed mathematically. 2) The edited object/scene conforms to the semantics of the text prompt. 3) The unedited parts of the object/scene retain the original identity.

Denoising diffusion models have emerged as a powerful tool for various image generation and editing tasks, facilitating the synthesis of visual content in an unconditional or input-conditional manner. The core idea behind them is learning to reverse the process of gradually adding noise to images, allowing them to generate high-quality samples from a complex distribution. In this survey, we provide an exhaustive overview of existing methods using diffusion models for image editing, covering both theoretical and practical aspects in the field. We delve into a thorough analysis and categorization of these works from multiple perspectives, including learning strategies, user-input conditions, and the array of specific editing tasks that can be accomplished. In addition, we pay special attention to image inpainting and outpainting, and explore both earlier traditional context-driven and current multimodal conditional methods, offering a comprehensive analysis of their methodologies. To further evaluate the performance of text-guided image editing algorithms, we propose a systematic benchmark, EditEval, featuring an innovative metric, LMM Score. Finally, we address current limitations and envision some potential directions for future research.

Amid the surge in generic text-to-video generation, the field of personalized human video generation has witnessed notable advancements, primarily concentrated on single-person scenarios. However, to our knowledge, the domain of two-person interactions, particularly in the context of martial arts combat, remains uncharted. We identify a significant gap: existing models for single-person dancing generation prove insufficient for capturing the subtleties and complexities of two engaged fighters, resulting in challenges such as identity confusion, anomalous limbs, and action mismatches. To address this, we introduce a pioneering new task, Personalized Martial Arts Combat Video Generation. Our approach, MagicFight, is specifically crafted to overcome these hurdles. Given this pioneering task, we face a lack of appropriate datasets. Thus, we generate a bespoke dataset using the game physics engine Unity, meticulously crafting a multitude of 3D characters, martial arts moves, and scenes designed to represent the diversity of combat. MagicFight refines and adapts existing models and strategies to generate high-fidelity two-person combat videos that maintain individual identities and ensure seamless, coherent action sequences, thereby laying the groundwork for future innovations in the realm of interactive video content creation.

Latest diffusion-based methods for many image restoration tasks outperform traditional models, but they encounter the long-time inference problem. To tackle it, this paper proposes a Wavelet-Based Diffusion Model (WaveDM). WaveDM learns the distribution of clean images in the wavelet domain conditioned on the wavelet spectrum of degraded images after wavelet transform, which is more time-saving in each step of sampling than modeling in the spatial domain. To ensure restoration performance, a unique training strategy is proposed where the low-frequency and high-frequency spectrums are learned using distinct modules. In addition, an Efficient Conditional Sampling (ECS) strategy is developed from experiments, which reduces the number of total sampling steps to around 5. Evaluations on twelve benchmark datasets including image raindrop removal, rain steaks removal, dehazing, defocus deblurring, demoiréing, and denoising demonstrate that WaveDM achieves state-of-the-art performance with the efficiency that is comparable to traditional one-pass methods and over 100× faster than existing image restoration methods using vanilla diffusion models. The code is available athttps://github.com/stayalive16/WaveDM

Neural radiance fields have made a remarkable breakthrough in the novel view synthesis task at the 3D static scene. However, for the 4D circumstance (e.g., dynamic scene), the performance of the existing method is still limited by the capacity of the neural network, typically in a multilayer perceptron network (MLP). In this article, we utilize 3D Voxel to model the 4D neural radiance field, short as V4D, where the 3D voxel has two formats. The first one is to regularly model the 3D space and then use the sampled local 3D feature with the time index to model the density field and the texture field by a tiny MLP. The second one is in look-up tables (LUTs) format that is for the pixel-level refinement, where the pseudo-surface produced by the volume rendering is utilized as the guidance information to learn a 2D pixel-level refinement mapping. The proposed LUTs-based refinement module achieves the performance gain with little computational cost and could serve as the plug-and-play module in the novel view synthesis task. Moreover, we propose a more effective conditional positional encoding toward the 4D data that achieves performance gain with negligible computational burdens. Extensive experiments demonstrate that the proposed method achieves state-of-the-art performance at a low computational cost.

Person Re-identification (ReID) aims to match people across non-overlapping camera views in a public space, which is usually regarded as an image retrieval problem to match query images with pedestrian images in the gallery. It is challenging since many difficulties exist such as pose misalignments, occlusions, similar appearance when detecting people. Existing researches on ReID mainly focus on two major problems: representation learning and metric learning. In this paper, we target at learning discriminative representations and make two contributions in total. (i) We propose a novel architecture named Deeply Associative Two-stage Representations Learning (DATRL). It contains the global re-initialization stage and fully-perceptual classification stage employing two identical CNNs associatively at the same time. On the global stage, we take on the backbone of one deep CNN e.g., dozens of layers in the front of Resnet-50 as a normal re-initialization subnetwork. Meanwhile, we apply our own proposed 3D-transpose technique into the backbone of the other CNN to form the 3D-transpose re-initialization subnetwork. The fully-perceptual stage is actually made up of the leftover layers of the original CNNs. On this stage, we take both the global representations learned at multiple hierarchies and the local representations uniformly-partitioned on the highest conv-layer into consideration, and then optimizing them separately for classification. (ii) We introduce a new joint loss function in which our proposed Labels Interval Extension loss (LIEL) and Group loss (GL) are combined to enhance the performance of gradient decent as well as increasing the distances between image features with different identities. We apply the above DATRL, LIEL and GL to ReID thus obtaining DATRL-ReID. Experimental results on four datasets CUHK03, Market-1501, DukeMTMC-reID and MSMT17-V2 demonstrate that DATRL-ReID shows excellent performance in improving recognition accuracy and is superior to state-of-the-art methods.

Developers can use different technologies for many software development tasks in their work. However, when faced with several technologies with comparable functionalities, it is not easy for developers to select the most appropriate one, as comparisons among technologies are time-consuming by trial and error. Instead, developers can resort to expert articles, read official documents or ask questions in QA sites for technology comparison, but it is opportunistic to get a comprehensive comparison as online information is often fragmented or contradictory. To overcome these limitations, we propose the diffTech system that exploits the crowdsourced discussions from Stack Overflow, and assists technology comparison with an informative summary of different comparison aspects. We first build a large database of comparable technologies in software engineering by mining tags in Stack Overflow, and then locate comparative sentences about comparable technologies with natural language processing methods. We further mine prominent comparison aspects by clustering similar comparative sentences and representing each cluster with its keywords. The evaluation demonstrates both the accuracy and usefulness of our model and we implement our approach into a practical website for public use.