arXiv:2506.02938v2 Announce Type: replace 
Abstract: Unsigned distance fields (UDFs) are widely used in 3D deep learning due to their ability to represent shapes with arbitrary topology. While prior work has largely focused on learning UDFs from point clouds or multi-view images, extracting meshes from UDFs remains challenging, as the learned fields rarely attain exact zero distances. A common workaround is to reconstruct signed distance fields (SDFs) locally from UDFs to enable surface extraction via Marching Cubes. However, this often introduces topological artifacts such as holes or spurious components. Moreover, local SDFs are inherently incapable of representing non-manifold geometry, leading to complete failure in such cases. To address this gap, we propose MIND (Material Interface from Non-manifold Distance fields), a novel algorithm for generating material interfaces directly from UDFs, enabling non-manifold mesh extraction from a global perspective. The core of our method lies in deriving a meaningful spatial partitioning from the UDF, where the target surface emerges as the interface between distinct regions. We begin by computing a two-signed local field to distinguish the two sides of manifold patches, and then extend this to a multi-labeled global field capable of separating all sides of a non-manifold structure. By combining this multi-labeled field with the input UDF, we construct material interfaces that support non-manifold mesh extraction via a multi-labeled Marching Cubes algorithm. Extensive experiments on UDFs generated from diverse data sources, including point cloud reconstruction, multi-view reconstruction, and medial axis transforms, demonstrate that our approach robustly handles complex non-manifold surfaces and significantly outperforms existing methods. The source code is available at https://github.com/jjjkkyz/MIND.

تتناول دراسة حديثة التحديات المتعلقة باستخراج الشبكات من حقول المسافة غير الموقعة (UDF) في التعلم العميق ثلاثي الأبعاد. على الرغم من أن حقول المسافة غير الموقعة قيمة لتمثيل الأشكال المعقدة، إلا أن تحقيق إعادة بناء دقيقة للأسطح أثبت أنه صعب. تبرز الأبحاث طريقة شائعة لتحويل حقول المسافة غير الموقعة إلى حقول المسافة الموقعة (SDF) لتسهيل هذه العملية. هذا العمل مهم لأنه يعالج عقبة رئيسية في النمذجة ثلاثية الأبعاد، مما قد يحسن التطبيقات في مجالات متنوعة مثل الألعاب، والواقع الافتراضي، ورسومات الكمبيوتر.

Un estudio reciente discute los desafíos de extraer mallas de campos de distancia no firmados (UDF) en el aprendizaje profundo 3D. Aunque los UDF son valiosos para representar formas complejas, lograr una reconstrucción de superficie precisa ha demostrado ser difícil. La investigación destaca un método común de convertir UDF en campos de distancia firmados (SDF) para facilitar este proceso. Este trabajo es significativo ya que aborda un obstáculo clave en la modelación 3D, lo que podría mejorar aplicaciones en diversos campos como los videojuegos, la realidad virtual y los gráficos por computadora.

Une étude récente aborde les défis liés à l'extraction de maillages à partir de champs de distance non signés (UDF) dans l'apprentissage profond 3D. Bien que les UDF soient précieux pour représenter des formes complexes, obtenir une reconstruction de surface précise s'est avéré difficile. La recherche met en lumière une méthode courante de conversion des UDF en champs de distance signés (SDF) pour faciliter ce processus. Ce travail est significatif car il traite un obstacle clé dans la modélisation 3D, pouvant améliorer les applications dans divers domaines tels que les jeux vidéo, la réalité virtuelle et les graphismes informatiques.

A recent study discusses the challenges of extracting meshes from unsigned distance fields (UDFs) in 3D deep learning. While UDFs are valuable for representing complex shapes, achieving precise surface reconstruction has proven difficult. The research highlights a common method of converting UDFs to signed distance fields (SDFs) to facilitate this process. This work is significant as it addresses a key hurdle in 3D modeling, potentially improving applications in various fields such as gaming, virtual reality, and computer graphics.

MIND: Material Interface Generation from UDFs for Non-Manifold Surface Reconstruction

Was this article worth reading? Share it

LucidQuery AI

Deptho.ai

Dyad

Z3D

Attentive AI

Artefacts.ai

Ready to build your own newsroom?