A Feature Selection Method for Ultrasound Image of Cervical Lymph Nodes

Abstract

The invention relates to a method for selecting features of ultrasound images of cervical lymph nodes, which relates to the technical field of image processing; the technical problem to be solved is to avoid the problem that the selection of features of ultrasound images of cervical lymph nodes by using a genetic algorithm and a discrete particle swarm optimization algorithm is easy to fall into a local extremum; The technical scheme adopted is as follows: the first step is to collect and extract the quantitative features of cervical lymph node images; the second step is to form a sample set; the third step is to design an extreme learning machine classifier; the fourth step is to learn and analyze the sample set Training; the fifth step, calculate the acceleration of each individual Step 6. Update their respective speeds The seventh step is to calculate the respective Boltzmann functions Step 8. Update their respective positions The ninth step, if the termination condition is met, then end and output the optimal position, that is, the optimal feature subset, otherwise the number of iterations is increased by 1, and return to the third step to continue the evolution of the population until the termination condition is met; the present invention is suitable for helping doctors select effective Value of ultrasound image features of cervical lymph nodes.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a feature selection method for lymph node ultrasound images based on a gravitational search algorithm and a Boltzmann function. Background technique [0002] Lymph nodes are distributed in the neck, armpits, groin and other parts of the entire human body. Whether the human immune system can function normally or not, lymph nodes play an important role. Clinically, pathological changes of lymph nodes are an important criterion for malignant tumors. Most clinicians use ultrasound images of cervical lymph nodes to diagnose and examine lymph node diseases. Ultrasound images of cervical lymph nodes have many lymph node characteristics, but some of these characteristics data are redundant. Therefore, selecting the features that are valuable to doctors is one of the focuses of clinicians and researchers. Feature selection is to select valuable features for classification, remov...

AI Technical Summary

Problems solved by technology

[0003] The present invention overcomes the deficiencies in the prior art, and the technical problem to be solved is to provide a feature selection method for ultrasonic images of cervical lymph nodes, which avoids the difficulty of using genetic algorithm and discrete particle swarm optimization algorithm for feature selection of ultrasonic images of cervical lymph nodes. The problem of getting stuck in a local extremum

Images