Medical Image Processor Recognition Frequency Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is a gap between the recognition of specific scenes in endoscopic images by a deep learning recognizer and the actual observation of these scenes by a user, leading to potential misinterpretation of notification indications.
Innovation Solution
A medical image processing apparatus that acquires medical images chronologically, recognizes specific scenes using deep learning, calculates a recognition frequency, and displays a notification indication on a monitor that changes in multiple stages to indicate the recognition frequency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a recognizer notifies the user that a specific scene has been observed, then the user is informed of the recognition result, but the user may wrongly recognize that the specific scene has been observed when it has not actually been observed
Solution Approach 1:
The system introduces feedback by displaying the recognition frequency of specific scenes to the user. The notification indication shows not only whether a scene is recognized but also how frequently it appears in the captured images, allowing users to verify the actual observation status and correct potential misrecognitions
Solution Approach 2:
The recognition frequency serves as an intermediary information between the recognizer's output and the user's perception. By inserting this intermediate metric, the system provides a more nuanced notification that bridges the gap between automated recognition and human observation verification
2Measurement precision
If the recognizer automatically recognizes specific scenes in endoscopic images, then recognition accuracy is improved, but a gap occurs between recognizer recognition and user observation
Solution Approach 1:
The system closes the observation gap by providing feedback on recognition frequency. Users can see how many times a recognized scene appears in the captured images, enabling them to verify whether the recognition corresponds to actual observation and reducing the information loss between automated and human recognition
Solution Approach 2:
The system adds a new dimension to the notification by incorporating frequency information. Instead of only indicating presence/absence of recognized scenes, the notification now includes quantitative frequency data, transforming a binary recognition state into a multi-dimensional information space that better reflects actual observation
Data Source
AI summary
Provided is a medical image processing apparatus, a medical image processing method, and a program that are capable of reducing a gap between recognition by a recognizer and recognition by a user. The medical image processing apparatus is a medical image processing apparatus including a processor (210), a memory (207), and a monitor (400). The processor (204) is configured to sequentially acquire a plurality of medical images in a chronological manner; recognize, on the basis of the acquired medical images, a specific scene in the medical images; acquire a recognition frequency of the recognized specific scene; and cause the monitor to display, in accordance with the recognition frequency, a notification indication that changes in two or more stages and that indicates a degree of recognition.