研究目的
To develop an automatic detection method of cell regions in microscope images based on bacterial foraging-based edge detection (BFED) algorithm for analyzing circulating tumor cells (CTC), addressing the challenges of manual analysis by pathologists.
研究成果
The proposed automatic detection method based on BFED algorithm achieved a true positive rate of 93.9% and false positive of 1.29 per case on 1680 images, reducing false positives compared to previous methods. Future work should focus on methods to emphasize only cell regions and improve robustness against brightness variations.
研究不足
The method is not robust to differences in image brightness due to varying shooting environments, leading to lower true positive rates in some cases. It also detects cell debris as false positives, and cannot handle regions where non-cell areas are fluorescently reacted.
1：Experimental Design and Method Selection:
The method involves three steps: segmentation using BFED algorithm for initial cell region detection, identification of single or multiple cells using SVM, and separation of connecting cells using branch and bound algorithm.
2：Sample Selection and Data Sources:
1680 red microscopy images from 6 cases, each with 1296x966 pixels, obtained using a fluorescence microscope.
3：List of Experimental Equipment and Materials:
DMi8 inverted fluorescence microscope (Leica, Germany), Universal CTC-Chip for sample preparation.
4：Experimental Procedures and Operational Workflow:
Input red images, extract salient regions using saliency map, create gradient map, apply BFED algorithm for segmentation, set ROIs, detect cell regions, identify cells with SVM, and separate cells with branch and bound algorithm.
5：Data Analysis Methods:
Performance evaluated using true positive rate (TPR) and false positive per image (FP), with comparison to a previous method.
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
