3D Vision

From Geometric Models to Deep Networks
Erscheinungsjahr 2027
ISBN: 978-981-9289-85-1
Verlag: Springer

Buch, Englisch, Format (B × H): 155 mm x 235 mm

The field of 3D vision is undergoing a major transformation, driven by the convergence of geometric modeling and deep learning. Classical geometric methods have long been central to 3D reconstruction, scene understanding, and object recognition, while data-driven approaches now open new pathways for tackling complex real-world challenges. Yet this shift has created a fragmented landscape where geometric rigor and learning-based flexibility often remain disconnected.

This academic monograph bridges that gap by offering a comprehensive and unified perspective on the evolution of 3D vision. It brings together diverse contributions from researchers and practitioners, presenting a curated selection of works that highlight innovation across theory, methodology, and application. Designed as both a scientific reference and a pedagogical resource, the volume provides structured overviews, case studies, benchmarks, and insights into practical implementations.

Intended for researchers, PhD students, engineers, and educators in computer vision, robotics, and artificial intelligence, the book combines theoretical depth with empirical results to document the paradigm shift in 3D vision. More than a collection of papers, it offers a cohesive narrative of how the field is evolving, equipping readers with tools, knowledge, and inspiration to engage with the future of 3D vision.

Zielgruppe

Research

Autoren/Hrsg.

Chambon, Sylvie

Quéau, Yvain

Fachgebiete

• Mathematik | Informatik EDV | Informatik Informatik Künstliche Intelligenz Computer Vision
• Mathematik | Informatik EDV | Informatik Informatik Bildsignalverarbeitung
• Mathematik | Informatik Mathematik Geometrie
• Technische Wissenschaften Elektronik | Nachrichtentechnik Elektronik

Inhaltsverzeichnis

Chapter 1. Paradigm Shift in 3D Vision.- Chapter 2. Inter-View Relationship in Light Field Image Quality Assessment for 3D Perception.- Chapter 3. Towards Data-Efficient Pixel-Wise Photometric Stereo.- Chapter 4. A Cross-Domain Distillation Study of Geometric Foundation Model for Stereo 3D Reconstruction.- Chapter 5. Self Supervised 3D from Monocular Videos.- Chapter 6. Multi-Scale Deformation-Guided Shape Matching with Unsupervised Deep Learning. Appendix: Activities of the working group.

Über Autor(innen)

Sylvie Chambon is a full professor at Toulouse INP, Institut National Polytechnique, affiliated with the IRIT laboratory in Toulouse, France. Her research lies at the intersection of image processing, computer vision, and deep learning, with applications in ecology, security, and healthcare. She has contributed to 14 international journal papers, like IJCV, Pattern Recognition or IEEE Transactions in Image Processing, and more than 40 international conference communications. She is the co-leader of the 3D Vision French working group.

Yvain Quéau is a permanent researcher (Chargé de Recherche) at the French National Centre for Scientific Research (CNRS), affiliated with the GREYC laboratory in Caen, France. His research focuses on the intersection of computer vision, geometric modeling, and machine learning, with a particular emphasis on 3D reconstruction, neural rendering, and hybrid approaches that combine geometric constraints with deep learning techniques. He has contributed to numerous peer-reviewed publications in top-tier conferences and journals, including CVPR, ICCV, IEEE TPAMI and IJCV. He has been the co-leader of the 3D Vision working group under the GdR IASIS (CNRS).