Medical Image Processing Apparatus for Endoscopic Image Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image processing apparatuses face challenges in accurately and efficiently selecting relevant medical images for diagnosis, particularly in endoscopic examinations, as they do not consider capturing information such as time, lesion details, treatment actions, or doctor's preferences, leading to an increased burden on doctors and potential selection of undesired images.
Innovation Solution
A medical image processing apparatus that acquires and analyzes imaging information, including accessory data like capturing time and position, to estimate the degree of interest for each image, classifying and selecting images based on this interest, using techniques like image processing to determine similarity, abnormality, and treatment state, thereby reducing the doctor's workload and improving selection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a predetermined number of medical images are selected from abnormal image groups using image processing, then the selection process is automated, but the selection accuracy and appropriateness decrease because capturing information and doctor's preferences are not considered
Solution Approach 1:
The patent changes the parameters used for image selection from simple image processing metrics to multiple parameters including capturing time, lesion information, treatment actions, and doctor's preferences. This allows the selection system to consider comprehensive factors while maintaining automation, thereby improving selection accuracy without reducing the extent of automation.
Solution Approach 2:
The patent implements feedback mechanisms where the system learns from doctor's selections and preferences over time. By incorporating feedback loops that adjust selection criteria based on actual doctor behavior and preferences, the system improves selection accuracy while maintaining automated operation.
2Reliability
If a large number of medical images are acquired in endoscopic examinations, then comprehensive diagnostic data is obtained, but the burden on doctors increases due to the difficulty of manual selection
Solution Approach 1:
The patent extracts and utilizes specific useful information from the large number of acquired images, such as capturing time, lesion information, and treatment actions. By extracting only the most relevant features and presenting a pre-selected subset of important images to doctors, the system maintains comprehensive diagnostic data while significantly reducing the doctor's selection burden.
Solution Approach 2:
The system performs self-service by automatically analyzing and prioritizing images based on embedded criteria (capturing time, lesion detection, treatment actions). This automated pre-processing and prioritization reduces the manual workload on doctors while ensuring comprehensive diagnostic information is preserved in the selected images.
3Ease of manufacture
If medical images are selected only from abnormal image groups, then the selection process is simplified, but the flexibility and adaptability decrease because the number of images to select varies according to doctor or hospital policy
Solution Approach 1:
The patent implements dynamic selection criteria that can adapt to different doctor and hospital policies. The system allows flexible configuration of selection parameters and can adjust the number and types of images selected based on varying policies, while maintaining a simplified automated process. This dynamic adaptability enables the system to accommodate different requirements without complicating the core selection mechanism.
Data Source
AI summary
An endoscopic image viewing support server includes an endoscopic image acquiring unit configured to acquire imaging information, a section-of-interest setting unit configured to estimate a degree of interest and classify a plurality of endoscopic images in accordance with the degree of interest, and an endoscopic image selecting unit configured to select an endoscopic image from each of sections of interest at a ratio based on the degree of interest. The section-of-interest setting unit is configured to determine an endoscopic image of interest from among the plurality of endoscopic images, and in a case where the plurality of endoscopic images are arranged in a chronological order, estimate the degree of interest for an endoscopic image in an endoscopic image group including the endoscopic image of interest by using image processing.