Abstract:Abstract In the undertaken study, a recognition system for circulating tumor cells was developed based on the YOLOv4 algorithm. The algorithm possesses the features of multi-layer high-level stacking, weight sharing, local connection and pooling that can be used to identify and category CTC tumor cells. The study is designed following five steps. The first step enriches many circulating tumor cells by targeting Negative Surface Charges of Cancer Cells by Multifunctional Nanoprobes. The second step screens the collected 1573 biological image samples based on circulating tumor cells according to the abnormal morphology of the nucleus and abnormal cytoplasm. The third step contains image preprocessing, unscrambling the training set and the test set, and training the YOLOv4 model along with improving the YOLOv4 model based on the test results. The fourth step has a circulating tumor cell feature extraction system based on the optimized YOLOv4 model to organize images of circulating tumor cells. The last step inputs the simulated blood sample cell image and the real blood sample cell image into the system for comparison and verification. When the proportion of the training set is 90% and the proportion of the test set is 10%, the model of IOU is the best, which is 98.88%. When the simulated blood sample image containing internal reference cells has put into the system, the average accuracy rate obtained from 10 experiments is 93.26%. When the real blood sample image has participated into the system, the average accuracy rate obtained from the 10 experiments is 85.67%. Comparing the current traditional pathology recognition, the artificial intelligence recognition system based on circulating tumor cells shows excellent computing performance and a lot time can be spared.

Deep Learning-Assisted Assessing of Single Circulating Tumor Cell Viability via Cellular Morphology

Circulating Tumor Cell Identification Based on Deep Learning

Antigen-Independent Single-Cell Circulating Tumor Cell Detection Using Deep-Learning-Assisted Biolasers

A Model for the Detection of Circulating Tumor Cells Based on Micro-OCT Imaging

Interpretatively Automated Identification of Circulating Tumor Cells from Human Peripheral Blood with High Performance

Digital Circulating Tumor Cells Quantification

A New Method for CTC Images Recognition Based on Machine Learning

Enrichment and Single-Cell Analysis of Circulating Tumor Cells.

Deep learning models for cancer stem cell detection: a brief review

Stain-free Holographic Detection of Circulating Tumor Cells Using A Deep Feature Fusion Neural Network

Single-cell Analyses of Circulating Tumor Cells

Automatic Detection of Circulating Tumor Cells with Very Deep Residual Networks

Evaluation of a Deep Learning Based Approach to Computational Label Free Cell Viability Quantification

Circulating Tumor Cells (CTCs): A Unique Model of Cancer Metastases and Non-invasive Biomarkers of Therapeutic Response

Recent progress on microfluidic technology-based circulating tumor cell analysis

Validation of a Microfluidic Device Prototype for Cancer Detection and Identification: Circulating Tumor Cells Classification Based on Cell Trajectory Analysis Leveraging Cell-Based Modeling and Machine Learning

Quantifying the Cell Morphology and Predicting Biological Behavior of Signet Ring Cell Carcinoma Using Deep Learning

Learning an Improved Object Detection Approach Based on the YOLO Algorithm to Identify Circulating Tumor Cells

Artificial Intelligence-driven Morphology-based Enrichment of Malignant Cells from Body Fluid

Machine Learning Assisted Dual-Modal SERS Detection for Circulating Tumor Cells

An Integrated Microfluidic Chip and Its Clinical Application for Circulating Tumor Cell Isolation and Single‐Cell Analysis