Extensive experiments show that HiFi-BRep significantly outperforms state-of-the-art methods in both structural validity and geometric fidelity, providing a robust solution for high-quality B-Rep synthesis.
Abstract
Boundary representation (B-Rep) generation is a fundamental task in computer-aided design, yet the direct synthesis of high-fidelity and structurally valid B-Reps remains a major challenge. Existing deep generative methods suffer from two forms of brittleness: representation brittleness, caused by padding noise and feature contamination in the latent space, and generation brittleness, stemming from sequential error propagation and a train-inference mismatch due to non-differentiable validity enforcement. We propose HiFi-BRep, a novel framework that addresses these limitations through two synergistic contributions. First, a topology-aware encoder constructs a high-fidelity latent representation by eliminating padding via learnable queries and preventing feature contamination with topology-guided attention. Second, a single-stage decoder jointly predicts geometry and topology in parallel, embedding core manifold constraints as a differentiable learning objective. This design ensures mutual guidance between geometry and topology while avoiding cascaded errors. Extensive experiments show that HiFi-BRep significantly outperforms state-of-the-art methods in both structural validity and geometric fidelity, providing a robust solution for high-quality B-Rep synthesis. Code and models are publicly available at https://github.com/1nnoh/HiFi-BRep.
This work introduces VAR-VTON, a VAR-based VTON model that incorporates garment conditioning and structural guidance for efficient latent-space VTON, and proposes STAR-VTON, a Two-Stage AutoRegressive framework that builds upon VAR-VTON by decoupling latent-space structural synthesis from pixel-space detail recovery.
This paper proposes Hybrid-Domain Posterior Sampling (HDPS), a decoupled inference framework that disentangles physical measurement consistency from semantic prior modeling, and establishes a new state-of-the-art, successfully recovering the high-frequency structural precision that latent-only solvers inherently discard.
Pose transfer, a core task in human-centric image generation, aims to synthesise photorealistic images of a subject in novel poses while preserving identity and intricate clothing details. Existing methods, particularly under large pose variations such as extreme articulation or self-occlusion, often struggle with preserving fine-grained textures and maintaining structural consistency, leading to artifacts like distorted limbs and lost details. To address these challenges, we introduce a unified generative adversarial network (GAN) framework that integrates three novel, complementary mechanisms. First, a Hierarchical Semantic Aligner (HSA) establishes multi-scale semantic correspondence between source appearance and target pose features through local attention and global gating. Second, a Pose-Aware Feature Injection (PAFI) module explicitly models source-target pose discrepancy to generate dynamic modulation parameters for adaptive feature adjustment during decoding. Third, a Gated Residual Fusion (GRF) strategy adaptively balances local detail and global structural information via a learnable dual-branch gating mechanism. Evaluated on the DeepFashion dataset, our framework demonstrates significant improvements, achieving a 13.9% reduction in Fréchet Inception Distance (FID) compared to the MAGPT method, alongside superior scores in Structural Similarity Index (SSIM) and Learned Perceptual Image Patch Similarity (LPIPS). Ablation studies confirm the individual contributions of each component. The proposed approach generates high-fidelity, structurally faithful results from a single reference image, offering a robust solution for applications in virtual try-on, animation, and human image synthesis under challenging pose transformations.
Tao Jiang, Guangyong Zheng, Songshui Wu et al.· IEEE Access· 0 citations
MotionCraft is presented, a controllable VSR framework that formulates restoration as motion-aware latent state prediction inspired by world models and integrates adaptive sparse attention with an explicit user-accessible control interface to deliver temporally consistent, high-quality reconstructions under streaming constraints.
Rong Fu, Chunlei Meng, Yangcheng Zeng et al.· 0 citations
Dataset distillation compresses a large training set into a compact synthetic set while retaining its downstream utility. Most existing methods target randomly initialized networks, whereas modern vision systems often adapt frozen pretrained encoders with lightweight modules. Distilled samples should therefore preserve the discriminative geometry of the pretrained representation space, which existing generative objectives do not explicitly consider. We propose self-supervised representation-guided generative dataset distillation (SRG), a framework that translates the SSL geometry into diffusion guidance. Specifically, SRG constructs class-wise prototypes from real-image SSL representations and performs guidance through three SSL-space objectives for prototype alignment, inter-class discrimination, and intra-class assignment. During diffusion sampling, it adopts a stage-wise guidance strategy: early denoising is anchored to the latent of the real image whose SSL representation is nearest to the assigned prototype, whereas later denoising is guided by the SSL-space objectives. This division preserves the visual realism provided by the generative prior while progressively steering samples toward representative and class-discriminative regions of the SSL representation space. SRG consistently outperforms the evaluated generative baselines across multiple datasets and IPC settings. A cross-encoder evaluation further indicates transfer across pretrained representation spaces. These results demonstrate the effectiveness of representation-guided generation for dataset distillation with pretrained SSL models.
Mingzhuo Li, Guangcheng Li, Linfeng Ye et al.· 0 citations
Text-to-motion generation aims to synthesize semantically consistent and naturally coherent motion sequences from natural language descriptions. Given the continuous nature of human motion, diffusion models operating in a continuous latent space offer inherent advantages over vector quantization-based methods, particularly in avoiding quantization errors and in modeling quality. However, existing diffusion models primarily rely on mean squared error loss. This stepwise regression paradigm often leads to ‘over-smoothed’ motion sequences and struggles to capture the subtle semantic nuances embedded in textual descriptions. To realize the potential for continuous diffusion generation, an enhanced latent-space diffusion framework designed to elevate generation capabilities across two dimensions, namely, distribution approximation and semantic alignment, is proposed. Specifically, a latent-space adversarial discriminator is incorporated. By applying decoupled adversarial supervision, this component mitigates the detail loss caused by mean regression, significantly enhancing the physical realism and dynamic sharpness. Concurrently, a latent-space contrastive alignment strategy is introduced during the denoising process that reinforces the correspondence of the generated motion sequences with the given textual inputs via explicit cross-modal constraints. Extensive experiments on standard benchmarks demonstrate that the proposed method effectively addresses the limitations of conventional diffusion models, thus validating the potential of continuous diffusion frameworks within the domain of text-driven motion synthesis.
Zhaowu Li, Rui Liu, Deheng Zhu et al.· Visual Computing for Industr...· 0 citations