Reciprocal Prosody Analysis for Real-Time Dominance Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods lack the ability to accurately and efficiently determine dominance behavior of participants in audio and video sessions in real-time or quasi-real-time, particularly in determining the relative dominance of individuals based on their audio interactions.
Innovation Solution
A system utilizing a processor and memory circuitry to extract features from audio content, generate baseline and updated baseline data, and employ a machine learning module to determine dominance using supervised learning, enabling real-time or quasi-real-time analysis of dominance behavior in multi-participant sessions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing methods are used to determine dominance behavior, then the analysis can be performed, but the accuracy and efficiency of real-time determination is insufficient
Solution Approach 1:
The patent segments the audio content into multiple time periods (initial period and subsequent periods) and extracts features for each segment separately. This allows the system to analyze dominance behavior at different time scales, improving both accuracy through detailed temporal analysis and efficiency through parallel processing of segments.
Solution Approach 2:
The patent extracts features and generates baseline data during an initial period before the main analysis. This preliminary action establishes a reference framework that enables more accurate and efficient dominance determination in subsequent periods, as the system already has baseline metrics to compare against.
2Adaptability or versatility
If baseline data and updated baseline data are extracted for multiple time periods, then the analysis can cover both short-term and long-term behavior, but the processing time and computational complexity increase
Solution Approach 1:
The patent divides the audio content into distinct time segments (initial period and multiple subsequent periods) and processes each segment independently. This segmentation enables the system to analyze both short-term behavior (within individual periods) and long-term behavior (across all periods) while managing processing time through structured, modular analysis.
Solution Approach 2:
The patent performs feature extraction and baseline data generation during the initial period as a preliminary step. This establishes a reference framework that reduces the computational burden in subsequent periods, as the system only needs to compare updated features against the pre-established baseline rather than analyzing everything from scratch.
3Extent of automation
If features are extracted from audio content and fed to machine learning module, then dominance can be determined automatically, but the system complexity increases
Solution Approach 1:
The patent introduces an intermediary feature extraction layer that bridges the audio content and the machine learning module. This intermediary component standardizes and prepares the audio features in a format suitable for machine learning processing, making the automatic determination process more manageable and reducing the apparent complexity by organizing the data flow in clear stages.
Solution Approach 2:
The patent segments the automated analysis process into distinct stages: feature extraction, baseline data generation, and machine learning inference. This segmentation makes the complex automated system more modular and easier to understand, as each stage has a clear, specific function that can be independently optimized and maintained.
Data Source
AI summary
A system comprising a processor and memory circuitry configured to, for at least one session comprising at least an audio content: for a first participant, extract features informative of an audio content associated with the first participant in a first initial period of time to generate first baseline data, for at least one first period of time starting after an end of the first initial period of time, extract features informative of an audio content associated with the first participant in the first period of time to generate first updated baseline data, perform similar operations for a second participant to generate second baseline data and second updated baseline data, feed the first and second baseline data, the first and second updated baseline data to a machine learning module to determine data informative of the dominance of the first participant and/or the second participant in the session.