Call for papers
Bridging Perception and Geometry: 3D Retrieval in the Era of Generative AI and Multimodal Understanding
Bridging Perception and Geometry: 3D Retrieval in the Era of Generative AI and Multimodal Understanding
About 3DOR
The International Symposium on 3D Object Retrieval (3DOR) is the premier venue for researchers and practitioners to share state-of-the-art methods, novel applications, and theoretical advancements in the field of 3D data management, searching, and retrieval.
Following the success of previous years, 3DOR 2026 seeks contributions that push the boundaries of current methodologies, particularly those exploring the intersection of deep learning, generative models, and holistic 3D understanding.
Topics of Interest
We invite high-quality submissions on all aspects of 3D object retrieval, similarity assessment, and understanding. Relevant topics include, but are not limited to:
Novel Retrieval Paradigms
Multimodal 3D Retrieval: Querying 3D models using text, sketches, images, audio, or other modalities.
Learning-Based Retrieval: Deep learning, contrastive learning, and self-supervised methods for 3D shape embeddings.
Generative Models for Retrieval: Using foundation models (e.g., LLMs/VLMs) for 3D retrieval and synthesis-by-query.
Category-Agnostic and Zero/Few-Shot Retrieval: Methods that generalize across object categories with limited data.
Semantic Segmentation and Scene Graph Generation in service of retrieval.
Core 3D Data Representation and Similarity
3D shape descriptors, feature extraction, and indexing techniques.
Geometry-based matching (point clouds, meshes, volumetric data).
Topology-based similarity and functional shape analysis.
Part-based retrieval and structural matching.
Handling of deformable, articulated, and non-rigid shapes.
Systems, Applications, and Evaluation
Interactive 3D search systems and user-centric evaluation metrics.
Benchmarking, datasets, and ground-truth creation for 3D retrieval.
Retrieval for specialized domains (e.g., medical imaging, cultural heritage, industrial CAD).
Large-scale 3D data management and efficient retrieval systems.
Real-time and lightweight retrieval algorithms for AR/VR applications.
Submission Guidelines
All papers must present original, unpublished work not currently under review elsewhere.
Format: Submissions must follow the standard C&G guide for authors (LaTeX recommended).
Length: Full papers should not exceed 12 pages (excluding references). Conference Proceedings (for ongoing work or novel applications) should not exceed 6 pages, unless the content of the paper warrants this.
Anonymity: We employ a double-blind review process. Authors must anonymize their submissions by removing all identifying information (names, affiliations, funding sources, etc.). Self-citations should be handled in the third person.
Publication: Full papers will be peer-reviewd by a two-stage review process, and published in a special section of Computers & Graphics (Elsevier). Proceedings papers will appear in the Eurographics Digital Library.