Embedded Dictation Detection for Physician Encounter Audio

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing systems struggle to accurately distinguish between dictation and non-dictation portions in audio recordings of physician-patient encounters, requiring physician input and leading to inefficiencies in scribing processes.

Innovation Solution

A system that utilizes ambient listening devices and machine learning algorithms to analyze speech patterns, proximity, and linguistic cues to automatically identify dictation segments, allowing for efficient navigation and transcription.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If existing systems rely on physician self-identification of dictation portions, then the system can capture all potential dictation segments, but the scribing efficiency decreases and physician workload increases

Engineering Contradiction:
Improvescribing efficiencyVSAvoidphysician time for self-identification
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The system enables self-service by allowing the audio recording to automatically identify and segment dictation portions through analysis of acoustic features, speech patterns, and contextual cues, eliminating the need for physician intervention in the identification process

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical process of manual physician identification with an automated computational system that uses machine learning models and audio analysis algorithms to detect and classify dictation segments automatically

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If the system analyzes only limited dictation portions, then processing speed increases, but measurement precision of dictation boundaries deteriorates

Engineering Contradiction:
Improvedictation boundary accuracyVSAvoidanalysis system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the audio recording into distinct segments based on detected dictation boundaries, allowing the system to process and analyze each segment independently with appropriate precision while managing overall system complexity through modular analysis

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system changes parameters by dynamically adjusting analysis thresholds, window sizes, and feature extraction parameters based on the detected speech characteristics and contextual information, enabling adaptive precision without fixed complex configurations

Inventive Principle:
Principle #35Parameter changes

3Productivity

If the scribe must listen to entire audio recordings linearly, then no information is missed, but the time required for transcription increases significantly

Engineering Contradiction:
Improvetranscription speedVSAvoidaudio playback time
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The system performs preliminary action by pre-identifying and marking all dictation segments before the scribe begins transcription, allowing the scribe to navigate directly to relevant portions rather than listening linearly through entire recordings

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces a temporal dimension by creating a non-linear navigation structure with time stamps and segment markers, allowing the scribe to jump between different portions of the recording based on the identified dictation segments rather than following linear playback

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP4233045B1Embedded dictation detection
Publication Date: 2025.11.26 SOLVENTUM INTELLECTUAL PROPERTIES CO

AI summary

Systems and techniques for identifying dictation within an audio stream are disclosed including training a first neural network model stored in a computer memory that is configured to process audio data to identify one or more audio segments that are indicative of dictation without using automatic speech recognition, training a second neural network model stored in the computer memory that is configured to process audio data using automatic speech recognition, analyzing the stored audio data using the first neural network to identify one or more segments in the audio data that are indicative of dictation, and analyzing the one or more identified segments that are indicative of dictation to extract one or more features from the one or more identified segments.