一种基于人工智能的染色体核型图像自动识别方法及系统
By acquiring the equidistant banding profile sequence of chromosome karyotype images and using a recurrent neural network, banding degradation curves are generated, solving the problem of distinguishing between enzymatically over-digested banding aberrations and true structural abnormalities, thus improving the accuracy and reliability of karyotype analysis.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HUNAN INST OF INFORMATION TECH
- Filing Date
- 2026-04-10
- Publication Date
- 2026-07-17
AI Technical Summary
Existing automatic image recognition technology has difficulty distinguishing between chromosome banding aberrations caused by excessive enzymatic digestion and true structural abnormalities, leading to distorted karyotype analysis results and increasing the risk of false positives in clinical diagnosis.
By acquiring chromosome karyotype images, the contours of equally spaced bands along the length direction are extracted, a sequence of band contours is constructed, and a pre-trained recurrent neural network is used to output band contrast, generate band degradation curves, and extract attenuation coefficients to determine whether it is an artificial aberration caused by excessive enzymatic digestion or a real structural abnormality.
It has achieved accurate identification of enzymatically over-digested band distortion, avoiding misjudgment, reducing the probability of false positives in clinical diagnosis, and improving the stability and intelligence level of automatic image recognition technology.
Smart Images
Figure CN122023399B_ABST