3D Eardrum Mapping via Deep Learning and Structured Light
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for reconstructing a three-dimensional map of an eardrum are cumbersome, requiring synchronization of projector and camera sequences, prone to motion artifacts, and often necessitate dyeing the eardrum, which increases procedure time and may cause side effects.
Innovation Solution
A computer-implemented method using a trained deep learning model to construct a three-dimensional map from a single two-dimensional representation of a deformed structured illumination pattern projected onto the eardrum, eliminating the need for sequential pattern projection and synchronization, and avoiding the need for dyeing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If temporal sequences of phase-shifted fringe patterns are projected onto the eardrum, then three-dimensional reconstruction accuracy is improved, but device complexity and synchronization requirements increase
Solution Approach 1:
The system pre-calculates and stores multiple fringe patterns with known phase shifts in a lookup table within the FPGA. Instead of dynamically generating and synchronizing phase-shifted sequences, the projector simply selects and displays the appropriate pre-prepared pattern from the lookup table based on the desired phase shift amount, eliminating complex real-time synchronization requirements while maintaining 3D reconstruction accuracy
Solution Approach 2:
The patent replaces the mechanical/electronic synchronization system with a software-based lookup table approach in the FPGA. The phase shift information is stored as data rather than being generated through coordinated hardware timing, substituting a complex synchronized control system with a simpler memory retrieval and display system
2Measurement precision
If multiple temporal sequences are projected and captured, then three-dimensional map accuracy is improved, but procedure time increases
Solution Approach 1:
Multiple fringe patterns with different phase shifts are pre-calculated and stored in the FPGA's lookup table before the measurement process begins. During the actual measurement, the system rapidly cycles through these pre-prepared patterns without needing to perform real-time phase shifting calculations or synchronize multiple temporal sequences, significantly reducing procedure time while maintaining the ability to reconstruct accurate three-dimensional maps
Solution Approach 2:
The system skips the time-consuming steps of real-time phase sequence generation and synchronization by using pre-computed patterns stored in memory. The projector rapidly displays the necessary patterns in sequence without waiting for synchronization signals, rushing through the measurement process efficiently while still capturing all necessary information for accurate 3D reconstruction
3Measurement precision
If the eardrum is dyed to increase visibility, then diagnostic accuracy is improved, but patient safety and procedure time worsen
Solution Approach 1:
The system changes the parameter of light intensity and uses high-contrast structured illumination patterns that can be clearly detected by the camera without requiring any chemical modification of the eardrum. By optimizing the optical parameters (light wavelength, intensity, pattern contrast), the system achieves high diagnostic accuracy through pure optical means, eliminating the need for harmful dyes
Solution Approach 2:
The patent substitutes the chemical method of dyeing the eardrum with an optical method using structured light projection and phase-shifting interferometry. Instead of using chemical substances to enhance visibility, the system uses carefully controlled light patterns and computational algorithms to extract three-dimensional information, replacing a potentially harmful chemical process with a safe optical measurement technique
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly reduces procedure time, enhances diagnostic accuracy by providing a true three-dimensional representation, and minimizes patient discomfort, allowing for efficient and precise eardrum examination without the need for phase-shifting or multi-shot setups.
Implementation Method 1
obtaining a two-dimensional representation of a reflection comprising a deformed illumination pattern of a structured illumination pattern projected onto the eardrum
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Example embodiments describe a method for obtaining a three-dimensional map of an eardrum comprising the steps of i) obtaining (702) a two-dimensional representation of a reflection comprising a deformed illumination pattern of a structured illumination pattern projected (701) onto the eardrum; and ii) constructing by a trained deep learning model the three-dimensional map based on the reflection. The deep learning model is further trained by a training dataset comprising a plurality of height maps and corresponding two-dimensional representations of a reflection comprising a deformed illumination pattern.