Breast Lesion Detection Model Using CNN Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for interpreting mammographic images for breast cancer diagnosis are time-consuming and prone to variability due to reliance on manual assessment and subjective interpretation by healthcare professionals, leading to inefficiencies and inaccuracies.
Innovation Solution
A diagnostic model and method using convolutional neural networks to classify and train mammographic images after processing and segmentation, enhancing the detection of breast lesion attributes such as location, margin, calcification, and size, thereby improving efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual assessment and subjective interpretation by healthcare professionals are used, then diagnostic accuracy can be maintained through expert judgment, but diagnostic time and personnel costs increase significantly
Solution Approach 1:
An automated image processing system acts as an intermediary between the mammographic images and the final diagnostic conclusion. The system performs preprocessing, segmentation, and feature extraction to generate standardized diagnostic reports, reducing the time burden on healthcare professionals while maintaining diagnostic accuracy through consistent application of diagnostic criteria
Solution Approach 2:
The manual mechanical process of image interpretation by healthcare professionals is replaced with an automated computational system. The system uses algorithms to automatically detect, segment, and characterize breast lesions, substituting human manual assessment with automated image analysis that operates faster and with consistent criteria
2Reliability
If manual interpretation of mammographic images is performed, then diagnostic experience and subjective perception are utilized, but judgment variability and inconsistency increase among different healthcare professionals
Solution Approach 1:
The system transforms subjective diagnostic parameters into objective, quantifiable measurements. By converting qualitative assessments into standardized numerical parameters and applying consistent algorithms, the system eliminates variability between different healthcare professionals while maintaining diagnostic reliability through reproducible results
3Measurement precision
If comprehensive image processing and segmentation are performed, then detection precision of breast lesion attributes is improved, but processing time and computational complexity increase
Solution Approach 1:
The image processing pipeline is divided into distinct sequential stages: preprocessing (denoising, contrast enhancement), segmentation (lesion identification and boundary detection), and feature extraction (characteristic measurement). This segmentation allows each stage to be optimized independently, achieving high detection precision while managing computational complexity through modular processing
Solution Approach 2:
Image preprocessing operations such as denoising and contrast enhancement are performed before segmentation and feature extraction. These preliminary actions prepare the images by reducing noise and enhancing lesion visibility, which simplifies subsequent processing steps and improves detection precision without requiring overly complex algorithms in later stages
Data Source
AI summary
A method is provided for building a model to determine breast lesions in a subject. The method involves sequential process of image processing, segmentation, object detection, and masking on obtained mammographic images to obtain local images and extracted feature information of breast lesions. Following this, classification and training are conducted using the local images and feature information to establish the model. Also provided herein is a method for diagnosing and treating breast cancer with the aid of the model.