Accurate polygonal building extraction requires not only region-level segmentation but also geometrically consistent vertex modeling. However, existing segment anything model (SAM)-based adaptations primarily optimize mask quality and lack explicit mechanisms for structured polygon prediction, leading to irregular contours and unstable vertex localization, especially under cross-domain scenarios. To address these limitations, we propose AsymSAM, an asymmetric feature-enhanced framework based on the SAM for polygonal building extraction. AsymSAM integrates an asymmetric hierarchical feature enhancement scheme and a hybrid fusion module in vanilla SAM. These components effectively leverage multilevel spatial and semantic information to enhance the representation of building contours and structural cues. Additionally, we extend the mask decoder to support joint prediction of polygon vertices and boundaries, enabling both automatic and prompt-guided extraction modes. Experiments on the WHU-Mix and SpaceNet datasets show AsymSAM consistently outperforms existing methods in terms of average precision (AP), average recall (AR), and intersection over union (IoU).

Synthesizing realistic 3D indoor scenes remains challenging due to data scarcity and the difficulty of simultaneously enforcing global architectural constraints and local semantic consistency. Existing approaches often overlook structural boundaries or rely on fully connected relation graphs that introduce redundant generation errors. Inspired by human design cognition, we present CasLayout, a cascaded diffusion framework that decomposes the joint scene generation task into four conditional sub-stages with explicit physical and semantic roles: (1) predicting furniture quantity and categories, (2) refining object sizes and feature embeddings, (3) modeling spatial relationships in a latent space, and (4) generating Oriented Bounding Boxes (OBBs). This decoupled architecture reduces data requirements and enables flexible integration of Large Language Models (LLMs) and Vision Language Models (VLMs) for zero-shot tasks such as image-to-scene generation. To maintain physical validity within complex floor plans, we explicitly model building elements (e.g., walls, doors, and windows) as conditional constraints. Furthermore, to address the high entropy of dense relation graphs, we introduce a sparse relation graph formulation aligned with human spatial descriptions. By encoding these sparse graphs into a compact latent space using a bidirectional Variational Autoencoder (VAE), the proposed framework provides enhanced relational controllability, allowing generated layouts to better respect functional organization. Experiments demonstrate that CasLayout achieves state-of-the-art performance in fidelity and diversity while enabling improved controllability in practical applications.

Components of the same category within facade images usually have similar semantic features and regular distribution patterns. Exploiting such structural priors can improve parsing model performance in information-deficient scenarios. We propose a pluggable heuristic strategy that enhances existing deep learning frameworks without architectural modifications, requiring only model-agnostic data processing and parameter-free feature map transformations. Our method introduces: (1) a deterministic pre-processing step using a simple segmentation model to evaluate the semantic clarity of input data, integrated with curriculum learning for stable parameter updates; and (2) latent-space feature transformations generating joint response maps to capture structural patterns. Experiments demonstrate that the method can significantly enhance the performance of host models, particularly when processing facade images with incomplete patterns caused by occlusions, low-light conditions, or similar challenges.

Point cloud normal estimation is a fundamental task in 3D geometry processing, playing a crucial role in applications such as 3D reconstruction, object recognition, and surface analysis. While recent learning-based methods achieve notable advancements in normal prediction, they often overlook the critical aspect of equivariance. This oversight leads to inefficient learning of symmetric patterns inherent in geometric data. To address this issue, we propose E$^{3}$-Net, an innovative neural network architecture designed to inherently achieve equivariance for normal estimation. We introduce an efficient random frame method, which significantly reduces the training resources required for this task to just 1/8 of previous work, while simultaneously enhancing prediction accuracy. Furthermore, we design a Gaussian-weighted loss function and a receptive-aware inference strategy that effectively leverage the local properties of point clouds, ensuring more precise and reliable normal estimation. Our method demonstrates superior performance across both synthetic and real-world datasets, consistently outperforming current state-of-the-art techniques by a substantial margin. Specifically, we achieve a 4% improvement in RMSE on the PCPNet dataset, 2.67% on the SceneNN dataset, and 2.44% on the FamousShape dataset, highlighting the robustness and scalability of E$^{3}$-Net in diverse environments.

Videos contain multi-modal content, and exploring multi-branch cross-modal interactions with natural language queries can be of benefit to the text-video retrieval task (TVR). However, recent methods applying the large-scale pre-trained CLIP model for TVR only focus on visual cues in videos. Furthermore, traditional methods of simply concatenating multimodal features do not exploit fine-grained cross-modal information in videos. In this paper, we propose a multi-branch multi-modal hybrid fusion (M2HF) network to hierarchically explore interaction between text queries and other modality content in videos. Specifically, M2HF first fuses visual features extracted by CLIP with audio and motion features extracted from videos to obtain fused audio-visual features and motion-visual features respectively. The multi-modal completion problem is also considered and solved in this process. Then, visual features, audio-visual features, motion-visual features, and text extracted from the video are used to establish cross-modal relationships with caption text queries using a multibranch approach. The retrieval outputs from all branches are then fused to obtain the final text-video retrieval results. Our framework provides two kinds of training strategies, using an ensemble approach and an end-to-end approach. Moreover, a novel multi-modal loss function is proposed to balance the contributions of each modality for efficient end-to-end training. M2HF allows us to obtain state-of-the-art results on various benchmarks: Rank@1 of 66.0%, 68.6%, 33.9%, 57.4%, and 57.3% on MSR-VTT, MSVD, LSMDC, DiDeMo, and ActivityNet, respectively.

The Medial Axis Transform (MAT) is a complete shape descriptor capable of reconstructing the geometry of the original domain. A high-quality MAT should not only facilitate high-fidelity reconstruction but also capture structural features – for instance, by aligning the MAT boundary with the locus of rolling ball centers within fillet regions. However, computing such an ideal MAT remains a significant challenge, particularly when the input is a discrete triangle mesh. In this paper, we follow the established technical pipeline of initializing the MAT via a 3D Voronoi diagram of surface samples and subsequently simplifying the Voronoi structure through a QEM-like scheme. Our key insight is to explicitly track the correspondence between MAT vertices and surface regions throughout the progressive simplification process, ensuring that the resulting MAT triangles accurately reflect the intrinsic symmetries between surface patches. We translate these geometric requirements into a suite of priority control strategies that govern the sequencing of edge collapses. Through extensive evaluation against state-of-the-art MAT algorithms, we validate the strong performance of our approach regarding runtime efficiency, structural alignment, boundary regularity, triangle quality, and robustness to noise. Our resulting MATs remain highly expressive for both articulated shapes and CAD models, even under extreme simplification – effectively capturing the global structure of complex geometries with only a few hundred vertices.

Creating robust, high-fidelity head avatars with 3D Gaussian Splatting (3DGS) remains constrained by two bottlenecks: one is the representation bottleneck that limits the modeling of dynamic details, and the other is the dynamic robustness bottleneck that causes information loss and point cloud inflation. To this end, we present LivingAvatars, which addresses both issues through two key designs: (i) a functionally decoupled Gaussian representation that separates identity-stable base Gaussians from detail Gaussians carrying high-frequency appearance, while a lightweight neural decoder injects features into the latter to synthesize expression-dependent details; and (ii) a Gaussian lifecycle management system that endows Gaussians with multiple states and replaces destructive pruning with probabilistic selective pruning and temporary hiding to suppress redundant growth, ensuring a compact and robust representation. On novel-view synthesis and self-reenactment, LivingAvatars achieves overall better metrics than existing methods. In cross-identity reenactment, it better preserves transient, expression-dependent signals, yielding higher perceptual fidelity. These results suggest that functionally decoupled Gaussian representations and lifecycle management are effective principles for building neural head avatars that remain robust and compact under dynamic occlusions, reliably reconstructing high-frequency details and improving overall realism.

View-based point cloud completion aims to recover a complete 3D shape from a partial point cloud, guided by a single-view image. However, existing approaches often suffer from limited performance due to weak modality alignment and limited self-geometry enhancement. To overcome these challenges, we propose a unified geometry-aware framework that integrates efficient modality alignment and adaptive geometry enhancement, mainly to address cross-modal geometric inconsistency of view-guided point cloud completion. Specifically, we propose a geometry-aware modality alignment by integrating a shared self-attention Transformer and cross-modality reconstruction supervision, which aims to bring features of the image and point cloud close to each other in a shared latent space describing the 3D object. To enhance the perception of global shape and local geometric details, we propose an adaptive geometry-aware self-attention module, which simultaneously considers local geometry-aware attention computation and the spatially-variant feature fusion. In addition, we apply a geometry-perceptive anchor refinement module to reorganize the anchor points (representing a local region of the shape) under appropriate supervision, further boosting the completion performance of our method. Extensive experiments on both synthetic and real-world datasets demonstrate that our method achieves superior performance over existing approaches. Our code will be available at https://github.com/weizequan/MAGE.

This is an erratum for the article “Loops2Roofs: Diffusion-based 3D Roof Generation using a Loop Representation” published in ACM Trans. Graph. 45, 5, Article 38 (June 2026), 16 pages.

Accurate polygonal building extraction requires not only region-level segmentation but also geometrically consistent vertex modeling. However, existing SAM-based adaptations primarily optimize mask quality and lack explicit mechanisms for structured polygon prediction, leading to irregular contours and unstable vertex localization, especially under cross-domain scenarios. To address these limitations, we propose AsymSAM, an asymmetric feature-enhanced framework based on the Segment Anything Model (SAM) for polygonal building extraction. AsymSAM integrates an asymmetric hierarchical feature enhance scheme and a hybrid fusion module in vanilla SAM. These components effectively leverage multi-level spatial and semantic information to enhance the representation of building contours and structural cues. Additionally, we extend the mask decoder to support joint prediction of polygon vertices and boundaries, enabling both automatic and prompt-guided extraction modes. Experiments on the WHU-Mix and SpaceNet datasets show AsymSAM consistently outperforms existing methods in terms of Average Precision (AP), Average Recall (AR), and Intersection over Union (IoU).

Recently, most existing point cloud frameworks tend to utilize max pooling aggregation functions to aggregate local point cloud features. However, when handling data containing local high-frequency noise such as local drop, addition, and jitter, this mechanism leads to high-frequency noise that spreads from local to global and causes severe performance degradation. To address this issue, we creatively extend the concepts from the physical field, namely electrostatic field and Coulomb forceinto geometric processing. To be specific, we treat the entire point cloud placed in an electrostatic field and each point as a probe charge and then equip this field with a set of source charges according to the structure of the cloud. We endow these two types of charges with different electric quantities, which could encode informative geometrical structural information. By analogously computing the Coulomb force between the probe charge and its corresponding source charge, we finally propose an explicit embedding called Point Geometric Coulomb Force (PGCF) for each point. Due to the deep use of the structural information of the point cloud and the fact that the electrostatic field of each source charge could not be affected by the variations of the probe charges, the PGCF has been proven to provide richer geometric information while being robust to local noises. Using the PGCF combined with point coordinates as inputs can significantly improve the performances of existing 3D point cloud feature extraction frameworks, including point convolution, graph convolution, and point transformer, without additional parameters or computational overhead, thus not affecting their inference speed. Experimental results show that integrating the PGCF into existing works brings more desirable results in a wide range of 3D point cloud analysis tasks, including classification, part segmentation, and semantic segmentation.

Composition is a high-level visual intent that governs where subjects are placed and how a scene is organized, yet current unified multimodal models remain unreliable at fine-grained composition recognition and struggle to turn such intent into controllable generation. We present COMPASS, the first unified multimodal framework that grounds composition-intent control in a single system spanning both composition perception and composition-guided generation, with a shared expert token τ_c as the central intent anchor. On the perception side, COMPASS injects composition expertise into an MoE backbone in a minimally invasive manner and distills the inferred intent into τ_c. On the generation side, COMPASS reuses τ_c as a global conditioning signal that steers the denoising trajectory, effectively converting passive composition analysis into explicit layout control. To support systematic instruction-following composition learning and evaluation at scale, we construct Comp-11, a large-scale dataset with an 11-class taxonomy and reasoning-augmented annotations. Extensive experiments show that COMPASS substantially improves category-level composition understanding and delivers more composition-consistent, prompt-faithful generation than strong baselines.

Videos contain multi-modal content, and exploring multi-branch cross-modal interactions with natural language queries can be of benefit to the textu2013video retrieval task (TVR). However, recent methods applying the large-scale pre-trained CLIP model for TVR only focus on visual cues in videos. Furthermore, traditional methods of simply concatenating multi-modal features do not exploit fine-grained cross-modal information in videos. In this paper, we propose a multi-branch multi-modal hybrid fusion (M2HF) network to hierarchically explore interaction between text queries and other modality content in videos. Specifically, M2HF first fuses visual features extracted by CLIP with audio and motion features extracted from videos to obtain fused audiou2013visual features and motionu2013visual features respectively. The multi-modal completion problem is also considered and solved in this process. Then, visual features, audiou2013visual features, motionu2013visual features, and text extracted from the video are used to establish cross-modal relationships with caption text queries using a multi-branch approach. The retrieval outputs from all branches are then fused to obtain the final textu2013video retrieval results. Our framework provides two kinds of training strategies, using an ensemble approach and an end-to-end approach. Moreover, a novel multi-modal loss function is proposed to balance the contributions of each modality for efficient end-to-end training. M2HF allows us to obtain state-of-the-art results on various benchmarks: Rank@1 of 66.0%, 68.6%, 33.9%, 57.4%, and 57.3% on MSR-VTT, MSVD, LSMDC, DiDeMo, and ActivityNet, respectively.

The reflective appearance, especially strong and typically near-field specular reflections, poses a fundamental challenge for accurate surface reconstruction and novel view synthesis. Existing Gaussian splatting methods either fail to model near-field specular reflections or rely on explicit ray tracing at substantial computational cost. We present Ref-DGS, a reflective dual Gaussian splatting framework that addresses this trade-off by decoupling surface reconstruction from specular reflection within an efficient rasterization-based pipeline. Ref-DGS introduces a dual Gaussian scene representation consisting of geometry Gaussians and complementary local reflection Gaussians that capture near-field specular interactions without explicit ray tracing, along with a global environment reflection field for modeling far-field specular reflections. To predict specular radiance, we further propose a lightweight, physically-aware specular adaptive mixing shader that fuses global and local specular features. Experiments demonstrate that Ref-DGS achieves state-of-the-art performance on reflective scenes while training substantially faster than ray-based Gaussian methods.

Video Moment Retrieval (MR) and Highlight Detection (HD) are closely related tasks that involve localizing semantically relevant video segments in untrimmed videos using natural language queries. Existing approaches struggle with the challenge of scattered relevant segments and varying durations, thereby resulting in limited localization performance. In this paper, we propose a Hierarchical Localization Framework (HLF) leveraging semantic relevance identification and hierarchical temporal perception to address both MR and HD tasks jointly. Specifically, we design a Video Feature Orientation (VFO) mechanism that explicitly computes token-level semantic importance by jointly considering attention distributions and cross-modal similarity cues. Subsequently, we employ a hierarchical information perception module designed to effectively capture diverse temporal patterns and leverage multi-scale features for prediction across varying timescales. Extensive experiments on several public benchmarks reveal that our method achieves superior performance on both MR and HD tasks.

Autoregressive models for 3D mesh generation suffer from a fundamental limitation: they flatten meshes into long vertex-coordinate sequences. This results in prohibitive computational costs, hindering the efficient synthesis of high-fidelity geometry. We argue this bottleneck stems from operating at the wrong semantic level. We introduce FACE, a novel Autoregressive Autoencoder (ARAE) framework that reconceptualizes the task by generating meshes at the face level. Our one-face-one-token strategy treats each triangle face, the fundamental building block of a mesh, as a single, unified token. This simple yet powerful design reduces the sequence length by a factor of nine, leading to an unprecedented compression ratio of 0.11, halving the previous state-of-the-art. This dramatic efficiency gain does not compromise quality; by pairing our face-level decoder with a powerful VecSet encoder, FACE achieves state-of-the-art reconstruction quality on standard benchmarks. The versatility of the learned latent space is further demonstrated by training a latent diffusion model that achieves high-fidelity, single-image-to-mesh generation. FACE provides a simple, scalable, and powerful paradigm that lowers the barrier to high-quality structured 3D content creation.

Recent advances in generative super-resolution (SR) have greatly improved visual realism, yet existing evaluation and optimization frameworks remain misaligned with human perception. Full-Reference and No-Reference metrics often fail to reflect perceptual preference, either penalizing semantically plausible details due to pixel misalignment or favoring visually sharp but inconsistent artifacts. Moreover, most SR methods rely on ground-truth (GT)-dependent distribution matching, which does not necessarily correspond to human judgments. In this work, we propose RefReward-SR, a low-resolution (LR) reference-aware reward model for preference-aligned SR. Instead of relying on GT supervision or NR evaluation, RefReward-SR assesses high-resolution (HR) reconstructions conditioned on their LR inputs, treating the LR image as a semantic anchor. Leveraging the visual-linguistic priors of a Multimodal Large Language Models (MLLM), it evaluates semantic consistency and plausibility in a reasoning-aware manner. To support this paradigm, we construct RefSR-18K, the first large-scale LR-conditioned preference dataset for SR, providing pairwise rankings based on LR-HR consistency and HR naturalness. We fine-tune the MLLM with Group Relative Policy Optimization (GRPO) using LR-conditioned ranking rewards, and further integrate GRPO into SR model training with RefReward-SR as the core reward signal for preference-aligned generation. Extensive experiments show that our framework achieves substantially better alignment with human judgments, producing reconstructions that preserve semantic consistency while enhancing perceptual plausibility and visual naturalness. Code, models, and datasets will be released upon paper acceptance.

Non-rigid alignment of point clouds is crucial for scene understanding, reconstruction, and various computer vision and robotics tasks. Recent advancements in implicit deformation networks for non-rigid registration have significantly reduced the reliance on large amounts of annotated training data. However, existing state-of-the-art methods still face challenges in handling occlusion scenarios. To address this issue, this paper introduces an innovative unsupervised method called Occlusion-Aware Registration (OAR) for non-rigidly aligning point clouds. The key innovation of our method lies in the utilization of the adaptive correntropy function as a localized similarity measure, enabling us to treat individual points distinctly. In contrast to previous approaches that solely minimize overall deviations between two shapes, we combine unsupervised implicit neural representations with the maximum correntropy criterion to optimize the deformation of unoccluded regions. This effectively avoids collapsed, tearing, and other physically implausible results. Moreover, we present a theoretical analysis and establish the relationship between the maximum correntropy criterion and the commonly used Chamfer distance, highlighting that the correntropy-induced metric can be served as a more universal measure for point cloud analysis. Additionally, we introduce locally linear reconstruction to ensure that regions lacking correspondences between shapes still undergo physically natural deformations. Our method achieves superior or competitive performance compared to existing approaches, particularly when dealing with occluded geometries. We also demonstrate the versatility of our method in challenging tasks such as large deformations, shape interpolation, and shape completion under occlusion disturbances.

A large number of 3D spectral descriptors have been proposed in the literature, which act as an essential component for 3D deformable shape matching and related applications. An outstanding descriptor should have desirable natures including high-level descriptive capacity, cheap storage, and robustness to a set of nuisances. It is, however, unclear which descriptors are more suitable for a particular application. This paper fills the gap by comprehensively evaluating nine state-of-the-art spectral descriptors on ten popular deformable shape datasets as well as perturbations such as mesh discretization, geometric noise, scale transformation, non-isometric setting, partiality, and topological noise. Our evaluated terms for a spectral descriptor cover four major concerns, i.e., distinctiveness, robustness, compactness, and computational efficiency. In the end, we present a summary of the overall performance and several interesting findings that can serve as guidance for the following researchers to construct a new spectral descriptor and choose an appropriate spectral feature in a particular application.