3D Scene Reconstruction, Generation and Understanding: Latest Advances and Prospects

Dear Colleagues,

In the digital era, 3D scene-related technologies have become pivotal in bridging the physical and virtual worlds, with applications spanning autonomous driving, augmented/virtual reality, robotics, and cultural heritage preservation. These technologies enable machines to perceive, model, and simulate the 3D environment, laying the foundation for intelligent interaction between humans, machines, and the world.

3D scene reconstruction focuses on accurately capturing the geometric information of real-world scenes from sensory data, which is essential for creating digital copies of real-world scenes. 3D scene generation involves creating realistic and plausible 3D scenes from scratch or based on given conditions, empowering the creation of virtual worlds with high fidelity. 3D scene understanding aims to interpret the semantic information, e.g., structure, objects, and relationships, within 3D scenes, enabling machines to comprehend the environment. Together, these three pillars—reconstruction, generation, and understanding—form the core of 3D scene analysis, driving innovation in numerous domains.

The aim of this Special Issue is to showcase the latest advances in these three interconnected areas and explore future prospects. Areas to be covered in this Special Issue may include recent and novel research trends related, but not limited, to the following:

• Novel algorithms for 3D scene reconstruction from single-/multi-modal data (e.g., RGB images, LiDAR, depth maps).
• 3D reconstruction algorithms for the challenges in low-quality data, such as occlusions, dynamic scenes.
• Efficient methods for large-scale and real-time 3D scene reconstruction.
• Semantic-aware 3D reconstruction techniques integrating object detection/segmentation.
• Neural network-based 3D scene generation, including conditional (text, sketches, etc.) and unconditional generation.
• Procedural and hybrid approaches for generating complex 3D scenes.
• Deep learning models for 3D scene understanding, such as object recognition, pose estimation, and relationship reasoning.
• Context-aware 3D scene parsing and semantic segmentation.
• Cross-modal understanding of 3D scenes (e.g., aligning 3D scenes with text).
• Benchmarks and evaluation metrics for 3D scene reconstruction, generation, and understanding.
• Applications of 3D scene technologies in specific domains (e.g., autonomous navigation, virtual try-on, digital preservation, and embodied AI).
• Ethical considerations and privacy issues in 3D scene data collection and usage.

The rapid developments in 3D scenes have had a profound impact on various domains, including embo. This Special Issue, entitled “3D Scene Reconstruction, Generation and Understanding: Latest Advances and Prospects”, aims to explore the integration of ML techniques in SE, with a particular emphasis on the transformative potential of Large Language Models (LLMs).

Dr. Qiankun Liu
Dr. Junbao Zhuo
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 250 words) can be sent to the Editorial Office for assessment.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Electronics is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2400 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

• NeRF
• 3D gaussian splatting
• point cloud
• 3D reconstruction
• 3D generation
• 3D understanding
• neural 3D modeling
• cross-modal scene analysis
• 3D scene benchmarks
• 3D scene applications