We present the Generalized Spatial Propagation Network (GSPN), a new attention mechanism optimized for vision tasks that inherently captures 2D spatial structures. Existing attention models, including transformers, linear attention, and state-space models like Mamba, process multi-dimensional data as 1D sequences, compromising spatial coherence and efficiency. GSPN overcomes these limitations by directly operating on spatially coherent image data and forming dense pairwise connections through a line-scan approach. Central to GSPN is the Stability-Context Condition, which ensures stable, context-aware propagation across 2D sequences and reduces the effective sequence length to $\sqrt{N}$ for a square map with N elements, significantly enhancing computational efficiency. With learnable, input-dependent weights and no reliance on positional embeddings, GSPN achieves superior spatial fidelity and state-of-the-art performance in vision tasks, including ImageNet classification, class-guided image generation, and text-to-image generation. Notably, GSPN accelerates SD-XL with softmax-attention by over $84\times$ when generating 16K images.
We present VARGPT, a novel multimodal large language model (MLLM) that unifies visual understanding and generation within a single autoregressive framework. VARGPT employs a next-token prediction paradigm for visual understanding and a next-scale prediction paradigm for visual autoregressive generation. VARGPT innovatively extends the LLaVA architecture, achieving efficient scale-wise autoregressive visual generation within MLLMs while seamlessly accommodating mixed-modal input and output within a single model framework. Our VARGPT undergoes a three-stage unified training process on specially curated datasets, comprising a pre-training phase and two mixed visual instruction-tuning phases. The unified training strategy are designed to achieve alignment between visual and textual features, enhance instruction following for both understanding and generation, and improve visual generation quality, respectively. Despite its LLAVA-based architecture for multimodel understanding, VARGPT significantly outperforms LLaVA-1.5 across various vision-centric benchmarks, such as visual question-answering and reasoning tasks. Notably, VARGPT naturally supports capabilities in autoregressive visual generation and instruction-to-image synthesis, showcasing its versatility in both visual understanding and generation tasks. Project page is at: \url{https://vargpt-1.github.io/}
Building on the success of diffusion models, significant advancements have been made in multimodal image generation tasks. Among these, human image generation has emerged as a promising technique, offering the potential to revolutionize the fashion design process. However, existing methods often focus solely on text-to-image or image reference-based human generation, which fails to satisfy the increasingly sophisticated demands. To address the limitations of flexibility and precision in human generation, we introduce ComposeAnyone, a controllable layout-to-human generation method with decoupled multimodal conditions. Specifically, our method allows decoupled control of any part in hand-drawn human layouts using text or reference images, seamlessly integrating them during the generation process. The hand-drawn layout, which utilizes color-blocked geometric shapes such as ellipses and rectangles, can be easily drawn, offering a more flexible and accessible way to define spatial layouts. Additionally, we introduce the ComposeHuman dataset, which provides decoupled text and reference image annotations for different components of each human image, enabling broader applications in human image generation tasks. Extensive experiments on multiple datasets demonstrate that ComposeAnyone generates human images with better alignment to given layouts, text descriptions, and reference images, showcasing its multi-task capability and controllability.
Racial bias in medicine, particularly in dermatology, presents significant ethical and clinical challenges. It often results from the underrepresentation of darker skin tones in training datasets for machine learning models. While efforts to address bias in dermatology have focused on improving dataset diversity and mitigating disparities in discriminative models, the impact of racial bias on generative models remains underexplored. Generative models, such as Variational Autoencoders (VAEs), are increasingly used in healthcare applications, yet their fairness across diverse skin tones is currently not well understood. In this study, we evaluate the fairness of generative models in clinical dermatology with respect to racial bias. For this purpose, we first train a VAE with a perceptual loss to generate and reconstruct high-quality skin images across different skin tones. We utilize the Fitzpatrick17k dataset to examine how racial bias influences the representation and performance of these models. Our findings indicate that the VAE is influenced by the diversity of skin tones in the training dataset, with better performance observed for lighter skin tones. Additionally, the uncertainty estimates produced by the VAE are ineffective in assessing the model's fairness. These results highlight the need for improved uncertainty quantification mechanisms to detect and address racial bias in generative models for trustworthy healthcare technologies.
Novel capacities of generative AI to analyze and generate cultural artifacts raise inevitable questions about the nature and value of artistic education and human expertise. Has AI already leveled the playing field between professional artists and laypeople, or do trained artistic expressive capacity, curation skills and experience instead enhance the ability to use these new tools? In this pre-registered study, we conduct experimental comparisons between 50 active artists and a demographically matched sample of laypeople. We designed two tasks to approximate artistic practice for testing their capabilities in both faithful and creative image creation: replicating a reference image, and moving as far away as possible from it. We developed a bespoke platform where participants used a modern text-to-image model to complete both tasks. We also collected and compared participants' sentiments towards AI. On average, artists produced more faithful and creative outputs than their lay counterparts, although only by a small margin. While AI may ease content creation, professional expertise is still valuable - even within the confined space of generative AI itself. Finally, we also explored how well an exemplary vision-capable large language model (GPT-4o) would complete the same tasks, if given the role of an image generation agent, and found it performed on par in copying but outperformed even artists in the creative task. The very best results were still produced by humans in both tasks. These outcomes highlight the importance of integrating artistic skills with AI training to prepare artists and other visual professionals for a technologically evolving landscape. We see a potential in collaborative synergy with generative AI, which could reshape creative industries and education in the arts.
Flow-matching models provide a powerful framework for various applications, offering efficient sampling and flexible probability path modeling. These models are characterized by flows with low curvature in learned generative trajectories, which results in reduced truncation error at each sampling step. To further reduce curvature, we propose block matching. This novel approach leverages label information to partition the data distribution into blocks and match them with a prior distribution parameterized using the same label information, thereby learning straighter flows. We demonstrate that the variance of the prior distribution can control the curvature upper bound of forward trajectories in flow-matching models. By designing flexible regularization strategies to adjust this variance, we achieve optimal generation performance, effectively balancing the trade-off between maintaining diversity in generated samples and minimizing numerical solver errors. Our results demonstrate competitive performance with models of the same parameter scale.Code is available at \url{https://github.com/wpp13749/block_flow}.
Diffusion models achieve superior performance in image generation tasks. However, it incurs significant computation overheads due to its iterative structure. To address these overheads, we analyze this iterative structure and observe that adjacent time steps in diffusion models exhibit high value similarity, leading to narrower differences between consecutive time steps. We adapt these characteristics to a quantized diffusion model and reveal that the majority of these differences can be represented with reduced bit-width, and even zero. Based on our observations, we propose the Ditto algorithm, a difference processing algorithm that leverages temporal similarity with quantization to enhance the efficiency of diffusion models. By exploiting the narrower differences and the distributive property of layer operations, it performs full bit-width operations for the initial time step and processes subsequent steps with temporal differences. In addition, Ditto execution flow optimization is designed to mitigate the memory overhead of temporal difference processing, further boosting the efficiency of the Ditto algorithm. We also design the Ditto hardware, a specialized hardware accelerator, fully exploiting the dynamic characteristics of the proposed algorithm. As a result, the Ditto hardware achieves up to 1.5x speedup and 17.74% energy saving compared to other accelerators.
Diffusion models have demonstrated exceptional capabilities in image generation and restoration, yet their application to video super-resolution faces significant challenges in maintaining both high fidelity and temporal consistency. We present DiffVSR, a diffusion-based framework for real-world video super-resolution that effectively addresses these challenges through key innovations. For intra-sequence coherence, we develop a multi-scale temporal attention module and temporal-enhanced VAE decoder that capture fine-grained motion details. To ensure inter-sequence stability, we introduce a noise rescheduling mechanism with an interweaved latent transition approach, which enhances temporal consistency without additional training overhead. We propose a progressive learning strategy that transitions from simple to complex degradations, enabling robust optimization despite limited high-quality video data. Extensive experiments demonstrate that DiffVSR delivers superior results in both visual quality and temporal consistency, setting a new performance standard in real-world video super-resolution.
Diffusion model shows remarkable potential on sparse-view computed tomography (SVCT) reconstruction. However, when a network is trained on a limited sample space, its generalization capability may be constrained, which degrades performance on unfamiliar data. For image generation tasks, this can lead to issues such as blurry details and inconsistencies between regions. To alleviate this problem, we propose a Sinogram-based Wavelet random decomposition And Random mask diffusion Model (SWARM) for SVCT reconstruction. Specifically, introducing a random mask strategy in the sinogram effectively expands the limited training sample space. This enables the model to learn a broader range of data distributions, enhancing its understanding and generalization of data uncertainty. In addition, applying a random training strategy to the high-frequency components of the sinogram wavelet enhances feature representation and improves the ability to capture details in different frequency bands, thereby improving performance and robustness. Two-stage iterative reconstruction method is adopted to ensure the global consistency of the reconstructed image while refining its details. Experimental results demonstrate that SWARM outperforms competing approaches in both quantitative and qualitative performance across various datasets.
Recent advances in text-driven image editing have been significant, yet the task of accurately evaluating these edited images continues to pose a considerable challenge. Different from the assessment of text-driven image generation, text-driven image editing is characterized by simultaneously conditioning on both text and a source image. The edited images often retain an intrinsic connection to the original image, which dynamically change with the semantics of the text. However, previous methods tend to solely focus on text-image alignment or have not aligned with human perception. In this work, we introduce the Text-driven Image Editing Benchmark suite (IE-Bench) to enhance the assessment of text-driven edited images. IE-Bench includes a database contains diverse source images, various editing prompts and the corresponding results different editing methods, and total 3,010 Mean Opinion Scores (MOS) provided by 25 human subjects. Furthermore, we introduce IE-QA, a multi-modality source-aware quality assessment method for text-driven image editing. To the best of our knowledge, IE-Bench offers the first IQA dataset and model tailored for text-driven image editing. Extensive experiments demonstrate IE-QA's superior subjective-alignments on the text-driven image editing task compared with previous metrics. We will make all related data and code available to the public.
Recent advancements in language-guided diffusion models for image editing are often bottle-necked by cumbersome prompt engineering to precisely articulate desired changes. An intuitive alternative calls on guidance from in-the-wild image exemplars to help users bring their imagined edits to life. Contemporary exemplar-based editing methods shy away from leveraging the rich latent space learnt by pre-existing large text-to-image (TTI) models and fall back on training with curated objective functions to achieve the task. Though somewhat effective, this demands significant computational resources and lacks compatibility with diverse base models and arbitrary exemplar count. On further investigation, we also find that these techniques restrict user control to only applying uniform global changes over the entire edited region. In this paper, we introduce a novel framework for progressive exemplar-driven editing with off-the-shelf diffusion models, dubbed PIXELS, to enable customization by providing granular control over edits, allowing adjustments at the pixel or region level. Our method operates solely during inference to facilitate imitative editing, enabling users to draw inspiration from a dynamic number of reference images, or multimodal prompts, and progressively incorporate all the desired changes without retraining or fine-tuning existing TTI models. This capability of fine-grained control opens up a range of new possibilities, including selective modification of individual objects and specifying gradual spatial changes. We demonstrate that PIXELS delivers high-quality edits efficiently, leading to a notable improvement in quantitative metrics as well as human evaluation. By making high-quality image editing more accessible, PIXELS has the potential to enable professional-grade edits to a wider audience with the ease of using any open-source image generation model.
Diffusion models represent a class of generative models that produce data by denoising a sample corrupted by white noise. Despite the success of diffusion models in computer vision, audio synthesis, and point cloud generation, so far they overlook inherent multiscale structures in data and have a slow generation process due to many iteration steps. In physics, the renormalization group offers a fundamental framework for linking different scales and giving an accurate coarse-grained model. Here we introduce a renormalization group-based diffusion model that leverages multiscale nature of data distributions for realizing a high-quality data generation. In the spirit of renormalization group procedures, we define a flow equation that progressively erases data information from fine-scale details to coarse-grained structures. Through reversing the renormalization group flows, our model is able to generate high-quality samples in a coarse-to-fine manner. We validate the versatility of the model through applications to protein structure prediction and image generation. Our model consistently outperforms conventional diffusion models across standard evaluation metrics, enhancing sample quality and/or accelerating sampling speed by an order of magnitude. The proposed method alleviates the need for data-dependent tuning of hyperparameters in the generative diffusion models, showing promise for systematically increasing sample efficiency based on the concept of the renormalization group.
In recent years, there has been an increasing interest in image anonymization, particularly focusing on the de-identification of faces and individuals. However, for self-driving applications, merely de-identifying faces and individuals might not provide sufficient privacy protection since street views like vehicles and buildings can still disclose locations, trajectories, and other sensitive information. Therefore, it remains crucial to extend anonymization techniques to street view images to fully preserve the privacy of users, pedestrians, and vehicles. In this paper, we propose a Street View Image Anonymization (SVIA) framework for self-driving applications. The SVIA framework consists of three integral components: a semantic segmenter to segment an input image into functional regions, an inpainter to generate alternatives to privacy-sensitive regions, and a harmonizer to seamlessly stitch modified regions to guarantee visual coherence. Compared to existing methods, SVIA achieves a much better trade-off between image generation quality and privacy protection, as evidenced by experimental results for five common metrics on two widely used public datasets.
In this project, we address the issue of infidelity in text-to-image generation, particularly for actions involving multiple objects. For this we build on top of the CONFORM framework which uses Contrastive Learning to improve the accuracy of the generated image for multiple objects. However the depiction of actions which involves multiple different object has still large room for improvement. To improve, we employ semantically hypergraphic contrastive adjacency learning, a comprehension of enhanced contrastive structure and "contrast but link" technique. We further amend Stable Diffusion's understanding of actions by InteractDiffusion. As evaluation metrics we use image-text similarity CLIP and TIFA. In addition, we conducted a user study. Our method shows promising results even with verbs that Stable Diffusion understands mediocrely. We then provide future directions by analyzing the results. Our codebase can be found on polybox under the link: https://polybox.ethz.ch/index.php/s/dJm3SWyRohUrFxn
Recent advancements in text-to-image generation models have excelled in creating diverse and realistic images. This success extends to food imagery, where various conditional inputs like cooking styles, ingredients, and recipes are utilized. However, a yet-unexplored challenge is generating a sequence of procedural images based on cooking steps from a recipe. This could enhance the cooking experience with visual guidance and possibly lead to an intelligent cooking simulation system. To fill this gap, we introduce a novel task called \textbf{cooking procedural image generation}. This task is inherently demanding, as it strives to create photo-realistic images that align with cooking steps while preserving sequential consistency. To collectively tackle these challenges, we present \textbf{CookingDiffusion}, a novel approach that leverages Stable Diffusion and three innovative Memory Nets to model procedural prompts. These prompts encompass text prompts (representing cooking steps), image prompts (corresponding to cooking images), and multi-modal prompts (mixing cooking steps and images), ensuring the consistent generation of cooking procedural images. To validate the effectiveness of our approach, we preprocess the YouCookII dataset, establishing a new benchmark. Our experimental results demonstrate that our model excels at generating high-quality cooking procedural images with remarkable consistency across sequential cooking steps, as measured by both the FID and the proposed Average Procedure Consistency metrics. Furthermore, CookingDiffusion demonstrates the ability to manipulate ingredients and cooking methods in a recipe. We will make our code, models, and dataset publicly accessible.
Generative models have made significant impacts across various domains, largely due to their ability to scale during training by increasing data, computational resources, and model size, a phenomenon characterized by the scaling laws. Recent research has begun to explore inference-time scaling behavior in Large Language Models (LLMs), revealing how performance can further improve with additional computation during inference. Unlike LLMs, diffusion models inherently possess the flexibility to adjust inference-time computation via the number of denoising steps, although the performance gains typically flatten after a few dozen. In this work, we explore the inference-time scaling behavior of diffusion models beyond increasing denoising steps and investigate how the generation performance can further improve with increased computation. Specifically, we consider a search problem aimed at identifying better noises for the diffusion sampling process. We structure the design space along two axes: the verifiers used to provide feedback, and the algorithms used to find better noise candidates. Through extensive experiments on class-conditioned and text-conditioned image generation benchmarks, our findings reveal that increasing inference-time compute leads to substantial improvements in the quality of samples generated by diffusion models, and with the complicated nature of images, combinations of the components in the framework can be specifically chosen to conform with different application scenario.
Recently, large-scale generative models have demonstrated outstanding text-to-image generation capabilities. However, generating high-fidelity personalized images with specific subjects still presents challenges, especially in cases involving multiple subjects. In this paper, we propose AnyStory, a unified approach for personalized subject generation. AnyStory not only achieves high-fidelity personalization for single subjects, but also for multiple subjects, without sacrificing subject fidelity. Specifically, AnyStory models the subject personalization problem in an "encode-then-route" manner. In the encoding step, AnyStory utilizes a universal and powerful image encoder, i.e., ReferenceNet, in conjunction with CLIP vision encoder to achieve high-fidelity encoding of subject features. In the routing step, AnyStory utilizes a decoupled instance-aware subject router to accurately perceive and predict the potential location of the corresponding subject in the latent space, and guide the injection of subject conditions. Detailed experimental results demonstrate the excellent performance of our method in retaining subject details, aligning text descriptions, and personalizing for multiple subjects. The project page is at https://aigcdesigngroup.github.io/AnyStory/ .
Throughout history, humans have created remarkable works of art, but artificial intelligence has only recently started to make strides in generating visually compelling art. Breakthroughs in the past few years have focused on using convolutional neural networks (CNNs) to separate and manipulate the content and style of images, applying texture synthesis techniques. Nevertheless, a number of current techniques continue to encounter obstacles, including lengthy processing times, restricted choices of style images, and the inability to modify the weight ratio of styles. We proposed a neural style transfer system that can add various artistic styles to a desired image to address these constraints allowing flexible adjustments to style weight ratios and reducing processing time. The system uses the VGG19 model for feature extraction, ensuring high-quality, flexible stylization without compromising content integrity.
Recently, LoRA and its variants have become the de facto strategy for training and sharing task-specific versions of large pretrained models, thanks to their efficiency and simplicity. However, the issue of copyright protection for LoRA weights, especially through watermark-based techniques, remains underexplored. To address this gap, we propose SEAL (SEcure wAtermarking on LoRA weights), the universal whitebox watermarking for LoRA. SEAL embeds a secret, non-trainable matrix between trainable LoRA weights, serving as a passport to claim ownership. SEAL then entangles the passport with the LoRA weights through training, without extra loss for entanglement, and distributes the finetuned weights after hiding the passport. When applying SEAL, we observed no performance degradation across commonsense reasoning, textual/visual instruction tuning, and text-to-image synthesis tasks. We demonstrate that SEAL is robust against a variety of known attacks: removal, obfuscation, and ambiguity attacks.
Large-scale text-to-image (T2I) diffusion models have demonstrated an outstanding performance in synthesizing diverse high-quality visuals from natural language text captions. Multiple layout-to-image models have been developed to control the generation process by utilizing a broad array of layouts such as segmentation maps, edges, and human keypoints. In this work, we present ObjectDiffusion, a model that takes inspirations from the top cutting-edge image generative frameworks to seamlessly condition T2I models with new bounding boxes capabilities. Specifically, we make substantial modifications to the network architecture introduced in ContorlNet to integrate it with the condition processing and injection techniques proposed in GLIGEN. ObjectDiffusion is initialized with pretraining parameters to leverage the generation knowledge obtained from training on large-scale datasets. We fine-tune ObjectDiffusion on the COCO2017 training dataset and evaluate it on the COCO2017 validation dataset. Our model achieves an AP$_{50}$ of 46.6, an AR of 44.5, and a FID of 19.8 outperforming the current SOTA model trained on open-source datasets in all of the three metrics. ObjectDiffusion demonstrates a distinctive capability in synthesizing diverse, high-quality, high-fidelity images that seamlessly conform to the semantic and spatial control layout. Evaluated in qualitative and quantitative tests, ObjectDiffusion exhibits remarkable grounding abilities on closed-set and open-set settings across a wide variety of contexts. The qualitative assessment verifies the ability of ObjectDiffusion to generate multiple objects of different sizes and locations.