Classification label construction method and device

Abstract

The embodiment of the invention provides a classification label construction method and device. The method comprises the steps of acquiring a sample picture used for training a preset model, wherein the preset model is a preliminary screening network model for identifying an original picture; wherein the sample picture is a second target picture containing abnormal characteristics on the outer surface of the shot object; classifying all the sample pictures according to the picture backgrounds of all the sample pictures; wherein the picture background comprises a picture background corresponding to a part of shooting defects and / or a picture background to which a covering is attached; comparing the classified sample pictures with each type of first target pictures respectively, classifyingthe classified sample pictures again according to a comparison result, and constructing labels in one-to-one correspondence for each type of classified sample pictures. The device executes the method.According to the classification label construction method and device provided by the embodiment of the invention, the reasonability and accuracy of classification label construction of the sample picture can be improved.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of image processing, and in particular, to a method and device for constructing classification labels. Background technique [0002] Capsule endoscopy has the advantages of being painless, non-invasive, and the captured images contain a large amount of information, and has a wide range of application values. [0003] The existing technology uses artificial methods to identify and classify the original pictures taken by capsule endoscopy. In order to identify the original pictures more accurately and efficiently, it is necessary to build a model, but the model usually needs to be trained before it is used. There is still no effective method for selecting training samples (which may be sample pictures) of pictures and constructing classification labels for the sample pictures. [0004] Therefore, how to avoid the above defects and improve the rationality and accuracy of the classificatio...

AI Technical Summary

Problems solved by technology

[0003] The existing technology uses artificial methods to identify and classify the original pictures taken by capsule endoscopy. In order to identify the original pictures more accurately and efficiently, it is necessary to build a model, but the model usually needs to be trained before it is used. There is no effective method for the selection of training samples (which can be sample pictures) of pictures and the construction of classification labels for sample pictures.

Images