Picture type cultural resource classification method
The invention discloses a picture type culture resource classification method, which comprises the following steps of: acquiring a plurality of types of culture resource image sets, dividing each type of image set into a training set and a test set, and preprocessing the training set; collecting tex...
Saved in:
Main Authors | , , , , , |
---|---|
Format | Patent |
Language | Chinese English |
Published |
19.05.2023
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The invention discloses a picture type culture resource classification method, which comprises the following steps of: acquiring a plurality of types of culture resource image sets, dividing each type of image set into a training set and a test set, and preprocessing the training set; collecting texture features of the images in the training set, and enabling each pixel point in the images to generate a corresponding LBP (Local Binary Pattern) code; according to the LBP codes, feature data corresponding to the subclass labels are formed, and the training set is used for training an identification model; performing texture feature extraction on the test set, inputting the texture features into the trained identification model, identifying local features, and calibrating subclass labels; carrying out K-means clustering on the images of which the texture features are extracted in the training set, constructing a BOW model, and obtaining a word bag required by the model; inputting the images in the test set into |
---|---|
Bibliography: | Application Number: CN202310152411 |