研究目的
To quantify lipids, collagen, and muscle cells in coronary arteries to help diagnose atherosclerosis using digital image processing techniques.
研究成果
The algorithm provides an objective method for quantifying substances in atherosclerotic plaques with a mean error less than 5% and processing time under two seconds, outperforming existing subjective methods. Future improvements could include dynamic thresholds for better accuracy.
研究不足
Errors in quantification are partly due to subjective expert assessments with a reported 25% margin of error. Difficulties in distinguishing muscle and foam cells from other substances lead to higher errors. The algorithm relies on fixed thresholds which may not adapt to all images, and processing times vary with region size.
1：Experimental Design and Method Selection:
The algorithm uses digital image processing techniques including object detection, color enhancement, shape and texture analysis, binarization, mode filtering, and color space conversion (e.g., to CIELab). It involves steps such as pre-processing, noise elimination, segment identification, lumen detection, interest region tracing, and quantification of specific substances.
2：Sample Selection and Data Sources:
Input images are artery sections dyed and captured by an electronic microscope at 20x magnification, in SVS format, with dimensions 0.125M x 0.125N. A sample of 50 medical images was used for evaluation.
3：125M x 125N. A sample of 50 medical images was used for evaluation.
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Electronic microscope (magnification 20x), computer for image processing, and software for implementing algorithms (specific brands/models not mentioned).
4：Experimental Procedures and Operational Workflow:
Steps include image input, binarization with thresholding, noise elimination using mode filters, segment identification with size thresholding, lumen detection using Moore-Neighbor Tracing algorithm, interest region tracing using gradient direction, and quantification of muscle cells, collagen, and foam cells through color and shape analysis in CIELab color space.
5：Data Analysis Methods:
Quantification is based on pixel counting in binary images. Error percentage is calculated by comparing algorithm results with expert evaluations using mean error percentage formula.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
