Detection method of pulmonary nodules based on three-dimensional convolution neural network
A neural network and three-dimensional convolution technology, applied in the field of medical image processing, can solve the problems of affecting the detection recall rate and not utilizing the essence of CT data, so as to improve the recall rate and average accuracy rate and reduce the time overhead
- Application Information
AI Technical Summary
Problems solved by technology
 Step 1: The origin and spacing in the image coordinate system are known during coordinate conversion, and the converted coordinates are calculated according to the following formula.
 Among them, center represents the position point to be converted, and center' represents the position point after conversion. Then, the threshold method and connected region labeling are used to obtain the labeled map of the lung parenchyma. Use morphological convex hull and expansion operations to repair the outer and inner contours of the lung parenchyma, obtain the mask of the lungs, and obtain the lung parenchyma area; then resample the data, and integrate the three surfaces in the patient coordinate system (cross-section , coronal plane, sagittal plane) were adjusted to 1 mm × 1 mm × 1 mm, and the value range of voxels was normalized to the range of 0-255 to obtain the preprocessing results of lung segmentation.
 Step 2: Use the defined network structure t...
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction