Personalized HRTF Prediction From Ear Images and Direction Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining personalized head-related transfer functions are cumbersome and time-consuming, requiring experimental measurements that are not practical for widespread application.
Innovation Solution
A method using an artificial neural network trained on images of ears and directional data to predict personalized head-related transfer functions, allowing for individualized sound experiences without the need for direct measurements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If experimental measurement methods are used to determine personalized head-related transfer functions, then measurement precision is improved, but device complexity and time consumption increase significantly
Solution Approach 1:
The patent creates a computational model that copies the physical measurement process through virtual simulation. Instead of performing actual acoustic measurements with microphones and loudspeakers, the system uses a trained neural network to predict HRTF values based on anthropometric data, effectively copying the measurement outcome without the physical measurement process itself
Solution Approach 2:
The patent replaces the mechanical measurement system (microphones, loudspeakers, anechoic chambers) with an information processing system. The neural network model substitutes physical acoustic measurements with computational predictions, transforming a mechanical/physical process into an information-based process that requires no physical measurement equipment
2Reliability
If experimental measurement procedures are implemented for personalized HRTF determination, then reliability of individualized sound experience is improved, but ease of operation deteriorates
Solution Approach 1:
The system copies the complex measurement procedure into a simplified data input process. Instead of requiring users to undergo time-consuming acoustic measurements, the system only requires input of anthropometric parameters that can be easily obtained from standard measurements or user profiles, maintaining reliability while dramatically improving ease of operation
3Manufacturing precision
If comprehensive training datasets with multiple users and directions are used, then manufacturing precision of the neural network model is improved, but quantity of data and processing requirements increase
Solution Approach 1:
The patent creates a universal neural network model that can predict HRTF for any user and any spatial direction simultaneously. The single trained model serves multiple functions: it handles different users with different anatomies, predicts HRTF for all 360 degrees of spatial directions, and works across different frequency ranges, eliminating the need for separate models for each user-direction combination
Solution Approach 2:
The patent transforms the training approach by adding the spatial direction dimension to the prediction task. Instead of training separate models for different directions, the system trains a single model that takes direction as an additional input parameter, allowing generalization across all spatial directions and reducing the effective data requirements
Data Source
AI summary
Computer-implemented method for determining a personalized head-related transfer function, the method comprising: receiving a training dataset comprising one or more training images of ears of one or more first users, a training input vector indicative of one or more first directions relative to the first user's head, and one or more values of the personalized head-related directional function related to the first users and the first directions; training an artificial neural network on the training dataset; receiving an inference dataset comprising an inference image of a second user's ear or a pair of inference images of the second user's ears, and an inference input vector indicative of a second direction relative to the second user's head; and processing the inference dataset by the artificial neural network to predict one or more personalized values of the directional function, the values related to the second user and the second direction.