We propose a compressive yet effective mesh tokenization, Blocked and Patchified Tokenization (BPT), facilitating the generation of meshes exceeding 8k faces. BPT compresses mesh sequences by employing block-wise indexing and patch aggregation, reducing their length by approximately 75% compared to the vanilla coordinate sequences. This compression milestone unlocks the potential to utilize mesh data with significantly more faces, thereby enhancing detail richness and improving generation robustness. Empowered with the BPT, we have built a foundation mesh generative model training on scaled mesh data to support flexible control for point clouds and images. Our model demonstrates the capability to generate meshes with intricate details and accurate topology, achieving SoTA performance on mesh generation and reaching the level for direct product usage.

Triangle meshes are fundamental to 3D applications, enabling efficient modification and rasterization while maintaining compatibility with standard rendering pipelines. However, current automatic mesh generation methods typically rely on intermediate representations that lack the continuous surface quality inherent to meshes. Converting these representations into meshes produces dense, suboptimal outputs. Although recent autoregressive approaches demonstrate promise in directly modeling mesh vertices and faces, they are constrained by the limitation in face count, scalability, and structural fidelity. To address these challenges, we propose Nautilus, a locality-aware autoencoder for artist-like mesh generation that leverages the local properties of manifold meshes to achieve structural fidelity and efficient representation. Our approach introduces a novel tokenization algorithm that preserves face proximity relationships and compresses sequence length through locally shared vertices and edges, enabling the generation of meshes with an unprecedented scale of up to 5,000 faces. Furthermore, we develop a Dual-stream Point Conditioner that provides multi-scale geometric guidance, ensuring global consistency and local structural fidelity by capturing fine-grained geometric features. Extensive experiments demonstrate that Nautilus significantly outperforms state-of-the-art methods in both fidelity and scalability. The project page is at https://nautilusmeshgen.github.io.

Generating compact and sharply detailed 3D meshes poses a significant challenge for current 3D generative models. Different from extracting dense meshes from neural representation, some recent works try to model the native mesh distribution (i.e., a set of triangles), which generates more compact results as humans crafted. However, due to the complexity and variety of mesh topology, most of these methods are typically limited to generating meshes with simple geometry. In this paper, we introduce a generic and scalable mesh generation framework PivotMesh, which makes an initial attempt to extend the native mesh generation to large-scale datasets. We employ a transformer-based autoencoder to encode meshes into discrete tokens and decode them from face level to vertex level hierarchically. Subsequently, to model the complex typology, our model first learns to generate pivot vertices as coarse mesh representation and then generate the complete mesh tokens with the same auto-regressive Transformer. This reduces the difficulty compared with directly modeling the mesh distribution and further improves the model controllability. PivotMesh demonstrates its versatility by effectively learning from both small datasets like Shapenet, and large-scale datasets like Objaverse and Objaverse-xl. Extensive experiments indicate that PivotMesh can generate compact and sharp 3D meshes across various categories, highlighting its great potential for native mesh modeling.

The next-coordinate prediction paradigm has emerged as the de facto standard in current auto-regressive mesh generation methods. Despite their effectiveness, there is no efficient measurement for the various tokenizers that serialize meshes into sequences. In this paper, we introduce a new metric Per-Token-Mesh-Entropy (PTME) to evaluate the existing mesh tokenizers theoretically without any training. Building upon PTME, we propose a plug-and-play tokenization technique called coordinate merging. It further improves the compression ratios of existing tokenizers by rearranging and merging the most frequent patterns of coordinates. Through experiments on various tokenization methods like MeshXL, MeshAnything V2, and Edgerunner, we further validate the performance of our method. We hope that the proposed PTME and coordinate merging can enhance the existing mesh tokenizers and guide the further development of native mesh generation.

Existing pretrained models for 3D mesh generation often suffer from data biases and produce low-quality results, while global reinforcement learning (RL) methods rely on object-level rewards that struggle to capture local structure details. To address these challenges, we present $\textbf{Mesh-RFT}$, a novel fine-grained reinforcement fine-tuning framework that employs Masked Direct Preference Optimization (M-DPO) to enable localized refinement via quality-aware face masking. To facilitate efficient quality evaluation, we introduce an objective topology-aware scoring system to evaluate geometric integrity and topological regularity at both object and face levels through two metrics: Boundary Edge Ratio (BER) and Topology Score (TS). By integrating these metrics into a fine-grained RL strategy, Mesh-RFT becomes the first method to optimize mesh quality at the granularity of individual faces, resolving localized errors while preserving global coherence. Experiment results show that our M-DPO approach reduces Hausdorff Distance (HD) by 24.6\% and improves Topology Score (TS) by 3.8\% over pre-trained models, while outperforming global DPO methods with a 17.4\% HD reduction and 4.9\% TS gain. These results demonstrate Mesh-RFT’s ability to improve geometric integrity and topological regularity, achieving new state-of-the-art performance in production-ready mesh generation.

The limitations of acquisition equipment often result in scene image data of limited size, posing a challenge for comprehensive analysis of social image datasets. Advances in generative models have introduced image outpainting techniques that expand the size of acquired social scene images, thereby enhancing the value of social image data. Stable diffusion (SD), which benefits from the guidance of caption prompts, shows excellent performance in image outpainting. However, its heavy reliance on manual prompts leads to a significant drawback: a decrease in the quality of generated images without prompts. To overcome this challenge, we propose a novel self-prompt diffusion model for image outpainting that extrapolates images based on the semantics of the source image, thereby removing the dependence on manual prompts. Specifically, we design a prompt autoencoder that uses an autoregressive transformer to map prompt embeddings into their semantic space, facilitating the construction of a semantic decoder. The semantic decoder and prompt embeddings are then cooptimized within the proposed prompt embedding network, allowing the mapping of image features to the stable diffusion prompt embeddings. Furthermore, by exploiting the inherent generative capabilities of diffusion models, we introduce a seam line regeneration mechanism to address the common problem of seam lines when splicing input and generated images. Comparative experiments on the Places2 and COCO datasets show that our method outperforms current state-of-the-art approaches on visual quality metrics and is adaptable to the stable diffusion model without additional fine-tuning.

Templates serve as a good starting point to implement a design (e.g., banner, slide) but it takes great effort from designers to manually create. In this paper, we present Desigen, an automatic template creation pipeline which generates background images as well as harmonious layout elements over the background. Different from natural images, a background image should preserve enough non-salient space for the overlaying layout elements. To equip existing advanced diffusion-based models with stronger spatial control, we propose two simple but effective techniques to constrain the saliency distribution and reduce the attention weight in desired regions during the background generation process. Then conditioned on the background, we synthesize the layout with a Transformer-based autoregressive generator. To achieve a more harmonious composition, we propose an iterative inference strategy to adjust the synthesized background and layout in multiple rounds. We constructed a design dataset with more than 40k advertisement banners to verify our approach. Extensive experiments demonstrate that the proposed pipeline generates high-quality templates comparable to human designers. More than a single-page design, we further show an application of presentation generation that outputs a set of theme-consistent slides. The data and code are available at https://whaohan.github.io/desigen.

Convolutional Neural Networks (CNNs) have ex-hibited great power in various vision tasks. However, the lack of transform-invariant property limits their further applications in complicated real-world scenarios. In this work, we pro-posed a novel generalized one-dimension convolutional operator (OneDConv), which dynamically transforms the convolution kernels based on the input features in a computationally and parametrically efficient manner. The proposed operator can extract the transform-invariant features naturally. It improves the robustness and generalization of convolution without sac-rificing the performance of common images. The proposed OneDConv operator can substitute the vanilla convolution. Thus, it can readily be incorporated into popular convolutional architectures, supporting end-to-end training. Empirical evaluations on popular benchmarks reveal OneDConv's superior performance over the standard convolution and competitive models in handling canonical and distorted images.

Image outpainting is a challenge for image processing since it needs to produce a big scenery image from a few patches. In general, two-stage frameworks are utilized to unpack complex tasks and complete them step-by-step. However, the time consumption caused by training two networks will hinder the method from adequately optimizing the parameters of networks with limited iterations. In this article, a broad generative network (BG-Net) for two-stage image outpainting is proposed. As a reconstruction network in the first stage, it can be quickly trained by utilizing ridge regression optimization. In the second stage, a seam line discriminator (SLD) is designed for transition smoothing, which greatly improves the quality of images. Compared with state-of-the-art image outpainting methods, the experimental results on the Wiki-Art and Place365 datasets show that the proposed method achieves the best results under evaluation metrics: the Fréchet inception distance (FID) and the kernel inception distance (KID). The proposed BG-Net has good reconstructive ability with faster training speed than those of deep learning-based networks. It reduces the overall training duration of the two-stage framework to the same level as the one-stage framework. Furthermore, the proposed method is adapted to image recurrent outpainting, demonstrating the powerful associative drawing capability of the model.

In the process of graphic layout generation, user specifications including element attributes and their relationships are commonly used to constrain the layouts (e.g.,"put the image above the button''). It is natural to encode spatial constraints between elements using a graph. This paper presents a two-stage generation framework: a spatial graph generator and a subsequent layout decoder which is conditioned on the previous output graph. Training the two highly dependent networks separately as in previous work, we observe that the graph generator generates out-of-distribution graphs with a high frequency, which are unseen to the layout decoder during training and thus leads to huge performance drop in inference. To coordinate the two networks more effectively, we propose a novel collaborative generation strategy to perform round-way knowledge transfer between the networks in both training and inference. Experiment results on three public datasets show that our model greatly benefits from the collaborative generation and has achieved the state-of-the-art performance. Furthermore, we conduct an in-depth analysis to better understand the effectiveness of graph condition modeling.