Patient-specific artificial neural network training system and method
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- SIEMENS HEALTHINEERS AG
- Filing Date
- 2023-07-20
- Publication Date
- 2026-07-21
AI Technical Summary
Existing artificial neural networks struggle to accurately diagnose patient-specific conditions due to insufficient or rare data, leading to misclassification and reduced diagnostic accuracy, particularly in medical applications like white blood cell differentiation.
A patient-specific training system that separates image data into reliable and unreliable sets based on classification confidence, training the neural network only with the reliable set, and refining the classification using patient-specific features.
Improves diagnostic accuracy by training the neural network with patient-specific data, enhancing its ability to classify images with high confidence and adapt to individual patient characteristics.
Smart Images

Figure 00000000_0000_ABST