Medical Image Processing Apparatus Textual Pattern Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image diagnostic techniques, such as those using X-ray CT images, rely heavily on human interpretation, which is burdensome and varies with doctor experience, and struggle to accurately extract effective feature values for computer-aided diagnosis due to direct extraction from images.
Innovation Solution
An image processing apparatus that acquires medical image data, calculates feature values based on spatial distributions of likelihood values representing textual patterns, using functions like likelihood acquisition and feature value calculation, incorporating methods like decision tree models and neural networks to enhance identification accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature values are directly extracted from a CT image, then the extraction process is simple, but it is difficult to extract feature values that are effective for identification
Solution Approach 1:
The patent introduces likelihood values as an intermediary between the CT image and the final feature values. The likelihood acquisition function computes likelihood values for each pixel indicating the probability of belonging to a specific textual pattern. These likelihood values then serve as the basis for calculating effective feature values, thereby mediating between raw image data and identification-ready features.
Solution Approach 2:
The patent replaces direct mechanical/extraction-based feature extraction from CT images with a computational approach using likelihood acquisition functions and decision tree models. Instead of directly extracting features from image pixels, the system substitutes this with computing likelihood values through trained models and then deriving features from these probabilities.
2Extent of automation
If computer-aided diagnosis is implemented with direct feature extraction from CT images, then automation is improved, but identification accuracy deteriorates
Solution Approach 1:
The patent performs preliminary action by acquiring likelihood values for each pixel before final feature extraction and identification. The likelihood acquisition function pre-computes probability distributions for each pixel belonging to different textual patterns, which then serves as refined input for subsequent feature calculation and machine learning-based identification, improving overall accuracy while maintaining automation.
3Measurement precision
If doctors perform image interpretation with naked eye, then detailed image interpretation is possible, but it is a significant burden on the doctor
Solution Approach 1:
The patent extracts the time-consuming and burdensome task of detailed image interpretation from doctors and transfers it to an automated system. The system extracts likelihood values for each pixel and computes feature values automatically, performing the detailed interpretation work that would otherwise require significant doctor time, while preserving the ability to provide detailed interpretation results.
Data Source
AI summary
According to one embodiment, an image processing apparatus includes processing circuitry. The processing circuitry is configured to acquire medical image data. The processing circuitry is configured to obtain spatial distribution of likelihood values representing a likelihood of corresponding to a textual pattern in a predetermined region of a medical image for each of a plurality of textual patterns based on the medical image data. The processing circuitry is configured to calculate feature values in the predetermined region of the medical image based on the spatial distribution obtained for the each of the plurality of textual patterns.