Endoscope System Audio Recognition Dictionary Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
During medical examinations using audio input with medical images, there is a risk of increased erroneous recognition and reduced operability due to the lack of appropriate information display, which existing techniques have not adequately addressed.
Innovation Solution
An endoscope system that includes an audio input device, an image sensor, and a processor, which acquires medical images chronologically, sets an audio recognition dictionary based on an audio input trigger, performs audio recognition using the set dictionary, and displays item information and recognition results on a display device, improving recognition accuracy and visual recognition for the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If audio recognition is performed on all words regardless of the scene, then the system can capture more information, but the accuracy of audio recognition decreases due to increased erroneous recognition between words
Solution Approach 1:
The patent applies local quality by setting different audio recognition dictionaries according to the examination scene. Instead of using a single universal dictionary for all words, the system selectively activates specific dictionaries (e.g., gastrointestinal tract dictionary, respiratory tract dictionary) based on the current examination context, thereby improving recognition accuracy for relevant terms while filtering out irrelevant words that could cause erroneous recognition.
Solution Approach 2:
The system changes the parameter of the audio recognition dictionary based on the examination scene. By dynamically switching between different dictionaries (changing the recognition parameter) according to the type of examination being performed, the system optimizes accuracy for each specific medical context while maintaining the ability to handle diverse examination scenarios.
2Loss of information
If the display device displays various kinds of information during the examination, then more information is available, but necessary information may not be displayed appropriately which hinders the examination procedure
Solution Approach 1:
The patent extracts and displays only the necessary audio recognition results related to the current examination scene on the display device. By filtering out irrelevant information and showing only the recognition results that are pertinent to the ongoing examination, the system maintains high information availability for necessary data while preventing information overload that would hinder the examination procedure.
Solution Approach 2:
The system performs preliminary filtering of audio recognition results based on the set dictionary before display. By pre-processing the recognition output to extract only relevant information matching the current examination context, the system ensures that the display shows only necessary information in advance, preventing clutter and maintaining procedural smoothness.
3Ease of operation
If a generic audio recognition system is used for all examinations, then the system is simpler to operate, but the recognition accuracy decreases for specific medical terminology in different examination contexts
Solution Approach 1:
The patent implements universality by creating a multi-functional audio recognition system that can handle multiple examination types through dictionary switching. The system maintains a universal framework that supports various specialized dictionaries (gastrointestinal, respiratory, etc.), allowing it to adapt to different medical contexts while preserving operational simplicity through automated dictionary selection based on the examination type.
Data Source
AI summary
According to the present disclosure, it is possible to smoothly proceed with an examination in which an audio input and audio recognition are performed on medical images. In the endoscope system, a processor acquires a plurality of medical images by causing an image sensor to image a subject in chronological order, accepts an input of an audio input trigger during capturing of the plurality of medical images, sets, in a case where the audio input trigger is input, an audio recognition dictionary according to the audio input trigger, performs, in a case where the audio recognition dictionary is set, audio recognition on audio input to the audio input device after the setting, using the set audio recognition dictionary, and displays item information indicating an item to be recognized using the audio recognition dictionary, and a result of audio recognition corresponding to the item information, on a display device.