Related Experiment Video
Updated: Jun 13, 2026
10:23
Three-Dimensional Cephalometric Landmark Annotation Demonstration on Human Cone Beam Computed Tomography Scans
Published on: September 8, 2023
An Annotation-Free Pipeline for 3D Auricular Bowl Atlas Construction and Statistical Shape Modelling from Surface
Tongxu Zhang1, Tony Kwok Wing Lee1, Jiebin Huang1
1Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Kowloon, Hong Kong, China.
Sensors (Basel, Switzerland)
|June 12, 2026
Summary
This study introduces an automated method for creating 3D ear models, crucial for designing hearing aids and medical devices. The new pipeline generates accurate statistical shape models without manual landmarks.
Area of Science:
- Biomedical Engineering
- Medical Imaging
- Computer-Aided Design
Background:
- Accurate 3D ear models are essential for medical devices like hearing aids and electrodes.
- Current methods often rely on time-consuming manual landmark placement.
- Developing automated pipelines is key for efficient ear shape modeling.
Purpose of the Study:
- To present a fully annotation-free pipeline for creating a 3D ear atlas and statistical shape model (SSM).
- To establish dense surface correspondences for 3D ear meshes.
- To provide a computational basis for optimizing ear-related devices.
Main Methods:
- Utilized 50 surface meshes to construct a 3D ear atlas and SSM.
- Employed iterative registration with the iterative closest point (ICP) algorithm and bidirectional thin-plate spline (BiTPS) deformation.
- Established dense correspondences via nearest-neighbor mapping and quantified registration quality.
Main Results:
- The BiTPS pipeline significantly reduced registration errors and increased coverage across five atlas iterations.
- SSM-derived dimensions showed excellent agreement with manual measurements (Pearson r≥0.98, ICC ≥0.98).
- The framework produced a stable, anatomically plausible ear atlas and an interpretable low-dimensional SSM.
Conclusions:
- The developed annotation-free pipeline successfully creates a 3D ear atlas and SSM.
- This method eliminates the need for manual landmarks, improving efficiency and accuracy.
- The framework supports geometric optimization for medical and wearable ear-related devices.
