Multi-View Image Processing System for Medical Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical imaging, particularly with mammograms, the presence of missing or duplicate images complicates the training of artificial intelligence models, as existing methods struggle to manage and utilize these images effectively, limiting the accuracy of Computer Aided Detection (CAD) systems.
Innovation Solution
The use of AI models, such as replacement AI models and quality generative AI models, to identify and generate replacement or improved images, addressing missing or duplicate images by training on time-adjacent sets of images, ensuring complete and high-quality datasets for training AI models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional image processing methods are used to handle missing or duplicate images, then the processing system is simple, but the accuracy of CAD systems is limited
Solution Approach 1:
The patent introduces AI models as intermediary components between the input images and the CAD system. These models process missing or duplicate images through sophisticated algorithms (including generative adversarial networks and attention mechanisms) to produce corrected image sets, thereby improving CAD accuracy without requiring changes to the core CAD system architecture
Solution Approach 2:
The system performs preliminary processing of images before they are fed into the CAD system. By detecting and correcting missing or duplicate images in advance using trained AI models, the system ensures that the CAD receiver receives clean, complete image data, thus improving diagnostic accuracy without adding complexity during the actual CAD operation
2Reliability
If AI models are used to generate replacement or improved images, then the quality of training datasets is improved, but the processing time and computational resources increase
Solution Approach 1:
The AI models are trained in advance on large datasets to learn the characteristics of medical images and the patterns of missing or duplicate images. This preliminary training allows the models to quickly process new images during actual use, reducing real-time processing time while maintaining high quality output
Solution Approach 2:
The system uses generative AI models to create synthetic replacement images that copy the essential features and characteristics of missing images. By generating realistic synthetic images rather than requiring actual physical images, the system improves dataset quality without the time and resource constraints of acquiring real images
Data Source
AI summary
One or more processors may identify a missing image in the set of multi-view images. Each of the images is associated with a particular view type. The one or more processors may generate, utilizing a replacement AI model, a replacement image for the missing image in the set of multi-view images. The replacement image is generated utilizing an AI model trained to generate a replacement image using training images from two or more time-adjacent sets of images. The one or more processors may identify a duplicate image in the set of multi-view images. The one or more processors may generate, utilizing a quality generative AI model, a characteristic improved image based on the duplicate image for the set of multi-view images. The one or more processors may output the replacement image and the characteristic improved image.