3D Lesion Volume Quantification in Biometric Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for evaluating bone tumor treatment and recurrence in 3D medical images rely on subjective interpretations, leading to inaccuracies due to the three-dimensional structure of bone tumors, necessitating a need for quantitative measurement and exact location of lesions to improve treatment evaluation and recurrence determination.
Innovation Solution
A computing device with a processor and memory executes steps to extract lesion information from 3D biometric images using a machine learning model, performs image processing to generate images highlighting the lesion region, and calculates the volume of the lesion quantitatively, including size and location, to enhance accuracy in treatment evaluation and recurrence determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods using 2D cross-section measurement are used, then the evaluation process is simple, but the measurement precision is poor due to the three-dimensional structure of bone tumors
Solution Approach 1:
The patent transitions from 2D cross-section measurement to 3D volumetric measurement by processing multiple axial CT images through machine learning segmentation. This dimensional transition enables accurate quantification of the three-dimensional tumor structure, resolving the contradiction between measurement precision and system complexity.
Solution Approach 2:
The patent replaces manual mechanical measurement methods with automated machine learning-based image processing. The deep learning model automatically segments and measures lesion volume from CT images, eliminating the need for manual 2D cross-section measurements and providing precise 3D quantification.
2Measurement precision
If qualitative interpretation of T1 and T2 relaxation characteristics is used, then the evaluation process is fast, but the measurement precision is poor for treatment evaluation and recurrence determination
Solution Approach 1:
The patent replaces subjective qualitative interpretation with automated machine learning-based quantitative measurement. The deep learning model objectively measures lesion volume and characteristics from CT images, providing precise treatment evaluation and recurrence determination without increasing evaluation time.
3Manufacturing precision
If manual measurement of long axis and short axis in 2D images is used, then the operation is simple, but the manufacturing precision (measurement accuracy) is insufficient for three-dimensional tumors
Solution Approach 1:
The patent replaces manual 2D measurement operations with automated 3D machine learning-based measurement. The deep learning model automatically segments the tumor in 3D space and calculates volumetric parameters, achieving high measurement accuracy while maintaining operational simplicity through automation.
Data Source
AI summary
Provided are a computing device and methods for quantifying a lesion in a biometric image. In certain aspects, disclosed a method including the steps of: extracting a lesion information in a plurality of first biometric images from each of the plurality of first biometric images three-dimensionally photographed of an object based on a machine learning model; generating a plurality of second biometric images in which a region of the lesion information, by performing image processing on each of the plurality of first biometric images; and calculating a volume of the lesion quantitatively using the region of the lesion information.