Snore Attribution Using AI Clustering and User Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for snore attribution in multi-person scenes require prior knowledge of the individuals and are not efficient when multiple people are sleeping in the same room, necessitating additional sampling steps.
Innovation Solution
A computer-implemented method using a trained model to distinguish snores of different individuals by forming vector space embeddings and clustering, combined with user input to attribute snores accurately without prior knowledge, utilizing a smartphone for local analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a trained model is used to disentangle snores of different people, then measurement precision of snore attribution is improved, but device complexity increases due to requiring a trained AI model and vector space embedding system
Solution Approach 1:
The patent segments the complex task of snore attribution into distinct processing stages: audio capture, snore detection, vector space embedding generation, clustering analysis, and user feedback integration. This segmentation allows each component to be optimized independently while maintaining overall system precision.
Solution Approach 2:
The patent introduces vector space embeddings as an intermediary representation that transforms raw audio data into a structured format suitable for clustering and comparison. This intermediary layer enables the system to handle complex attribution tasks without requiring direct complex processing of raw audio signals.
2Measurement precision
If user feedback is collected through a user interface for each played snore, then measurement precision of snore attribution is improved, but loss of time increases due to requiring user interaction for each snore
Solution Approach 1:
The patent applies partial action by selecting and playing only a subset of snores for user feedback rather than requiring feedback on all detected snores. This reduces the time burden on users while still providing sufficient data for accurate attribution through the clustering algorithm.
Solution Approach 2:
The system implements a feedback mechanism where user responses to played snores are used to refine and update the clustering model. This feedback loop improves attribution accuracy over time without requiring continuous user interaction for every snore event.
3Ease of operation
If prior knowledge of individuals is not required, then ease of operation is improved, but measurement precision may worsen due to lack of pre-sampled characteristics
Solution Approach 1:
The patent performs preliminary action by automatically generating vector space embeddings from raw audio data during the initial detection phase. This pre-processing step creates structured representations that enable accurate clustering and attribution without requiring prior knowledge or manual sampling of individual characteristics.
Solution Approach 2:
The system implements self-service by automatically adapting to different individuals through unsupervised clustering algorithms that learn snore patterns directly from the audio data. The model self-adjusts to identify and attribute snores to different individuals without requiring pre-registration or prior knowledge of the users.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present invention relates to a computer-implemented method for snore attribution, comprising: capturing audio using a microphone (12); detecting a plurality of snores in the captured audio; determining that a first set of snores (42a) belongs to a first individual (28a) and that a second set of snores (42b) belongs to a second individual (28b) using a trained model (44); playing for a user a subset of the snores of said first set (42a) and a subset of the snores of said second set (42b); for each played snore, prompting the user via a user interface (18) to provide input whether or not the played snore belongs to the user; receiving said input from the user via the user interface; and attributing the first set of snores or the second set of snores to the user based on said input.