End-to-end blood cell recognition model construction method and application

Abstract

The invention relates to an end-to-end blood cell recognition model construction method and application. A data sample set is formed based on a full-view image. An artificial intelligence technology is used for training a blood cell recognition model, and a mature recognition model is finally formed through continuous parameter optimization and error analysis optimization of the model.The model input is a single-view blood smear image, and the output is all cell positions, edges and categories on the image. Full-view blood cell analysis is achieved through a computer, interference of human objective factors is greatly reduced, and objectivity and consistency of inspection results are improved. The blood cell recognition model is intelligent, the software algorithm has a self-learning attribute, the training efficiency of the recognition model is gradually improved along with the increase of high-quality labeled images, and the software recognition and classification accuracy can be continuously optimized.

Description

technical field [0001] The invention relates to an end-to-end blood cell recognition model construction method and its application, belonging to the technical field of medical images. Background technique [0002] The current blood test process in the hospital is: blood sample - blood analyzer - push dye machine - manual microscopic examination, the whole process takes about 60 minutes. Blood samples are obtained by manual blood drawing; various blood cell counts, white blood cell classification and hemoglobin content are obtained through blood analysis instruments; dyeing and marking are carried out through push-staining machine to slides for manual microscopic examination; final blood cell morphology analysis results are obtained after manual microscopic examination , such as identification of abnormal blood cells, etc. [0003] Existing hematology analyzer technologies are mainly based on electrical impedance, laser measurement and comprehensive methods (flow cytometry, ...

AI Technical Summary

Problems solved by technology

[0004] The problems of the prior art are: first, the analysis and counting of blood cells in the full field of view of the blood smear are not realized, and the data sample size is not enough, resulting in one-sided and inaccurate results; second, the counting and classification algorithms are relatively traditional, and the effect of morphological analysis is relatively low. Poor, the recognition accuracy is not high; Third, there is a serious shortage of high-level medical examiners, and the subjectivity of manual microscopic examination doctors cannot be controlled, and the re-examination rate is high; Fourth, the time is long and the efficiency is low

Images