Tumor prediction method based on image feature data

Abstract

The invention discloses a tumor prediction method based on image feature data. The method comprises steps that S1 of reference tumor data collection, collection of tumor image data and historical tumor data of different types and degrees of tumor patients is performed; S2 of establishment of a reference tumor image feature database and a historical tumor database, the tumor image data of the tumorpatients described in the S1 and the historical tumor data are classified, and the reference tumor image feature database and the historical tumor database are established; and S3 of establishment ofa contrast function, based on the reference tumor image feature database and the historical tumor database of the S2, through a Kalman filter, a finite element and the neural network algorithm, the comparison function is established. The method is advantaged in that the reference tumor image feature database and the historical tumor database are established through collecting the tumor image dataof the tumor patients, a comparison function model is established through classification, tumor prediction can be accurately performed, the comparison function model can be well revised through observing the late incidence of subjects, so the tumor can be predicted better, and people are provided with security.

Description

technical field [0001] The invention relates to the technical field of tumor diagnosis, in particular to a tumor prediction method based on image feature data. Background technique [0002] Tumor tissue differs from the normal tissue from which it originates in varying degrees in terms of cell morphology and tissue structure, and this difference is called atypia. Atypia is the morphological manifestation of abnormal differentiation of tumors. A small atypia indicates a high degree of differentiation, and a large atypia indicates a low degree of differentiation. Distinguishing the size of this atypia is the main histological basis for diagnosing tumors and determining benign and malignant tumors. The atypia of benign tumor cells is not obvious and is generally similar to its source tissue. Malignant tumors often have obvious atypia. [0003] Benign tumors do little harm to the human body, while malignant tumors do great harm to the human body. Malignant tumors are divided...

AI Technical Summary

Problems solved by technology

The present invention establishes a database by collecting tumor image data of patients with different types of malignant tumors, and then establishes a proportional hazard function, then converts the data of the subject and fills in missing values, and then substitutes the filled values ​​into the proportional hazard function to establish a tumor The prediction model can predict tumors more accurately, but it only evaluates the overall data in general, so as to calculate the risk value of the subject's disease. The accuracy is low, and it cannot revise the function better. Therefore, we urgently need a tumor prediction method based on image feature data to solve the above problems.