PROCEDURE FOR AUTOMATED CHROMOSOME ANALYSIS
Patent Information
- Application Number
- DE602020053432
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2019-10-17
- Filing Date
- 2020-10-19
- Publication Date
- 2025-06-25
- Estimated Expiration
- 2040-10-19
AI Technical Summary
Current chromosome analysis methods, particularly karyotyping, face challenges in accuracy and efficiency due to the need for preparatory processing steps and user intervention, especially in cases of poor banding quality and high error rates in automated recognition, making it difficult to streamline the diagnostic process.
The application of deep convolutional neural networks (DNNs) for chromosome segmentation and classification, reducing the need for preparatory processing and user intervention, and providing precise orientation and classification of chromosomes.
The DNN-based approach significantly reduces user interaction and improves classification accuracy, achieving a 98.3% overall accuracy and a 25% reduction in interactive corrections, enhancing the efficiency and reliability of chromosome analysis.