The invention provides a nasal polyp automatic grading model training method, a scoring method, a
system and equipment, and relates to the technical field of
image analysis. In order to solve the problem that automatic grading of nasal polyp of a
chronic sinusitis patient is difficult to realize in the prior art, an
endoscopic image set containing different grades of nasal polyp
score marking information is constructed based on nasal
endoscopic image data of the patient, and a preset target detection model is trained by using the image set, so that the classification of the nasal polyp of the
chronic sinusitis patient is realized. Therefore, the preset target detection model can directly learn and internalize the
international standard of nasal polyp grade scoring, and then the preset target detection model can stably learn the
nasal endoscope characteristics covering each scoring grade through the full-
process training of data enhancement,
cross validation and hyper-parameter tuning, so that the accuracy of the nasal
endoscopy is improved. And the output nasal polyp grade
score has both
automation and accuracy, so that the core technical
bottleneck of automatic nasal polyp grading under the
endoscope is solved.