Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Terracotta army fragment classification method based on deep learning

A technology of deep learning and classification methods, applied in neural learning methods, instruments, biological neural network models, etc., can solve the problems of inaccurate classification, low measurement accuracy, and high degree of experience dependence, and achieve the effect of convenient operation

Pending Publication Date: 2021-03-02
NORTHWEST UNIV(CN)
View PDF5 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a method for classifying fragments of Terracotta Warriors based on deep learning to solve the problems of high dependence on experience, low measurement accuracy and inaccurate classification in the prior art

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Terracotta army fragment classification method based on deep learning
  • Terracotta army fragment classification method based on deep learning
  • Terracotta army fragment classification method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0063] Such as figure 1 As shown, the fragment data set of the Terracotta Warriors in this embodiment is the fragment model collected by the researchers of the National and Local Joint Engineering Research Center of Cultural Heritage Digitalization of Northwest University to the Museum of Terracotta Warriors and Horses. There are more than 500 existing fragment models, and a large Some have been accurately labeled. Divide the fragments into 6 body parts, including arms, feet, skirts, upper body, hands and legs, as 6 categories, and capture the images of the outer surface of the fragments. Each fragment model has about 10 screenshots, a total of 2000 Multiple pictures.

[0064] The following are the specific steps:

[0065] Step 1: Use python to preprocess the image of the terracotta warriors, remove the background of the picture, cut off the redundant white edges, and then unify the RGB image with a picture size of 227*227, and perform normalization processing to obtain 1800...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a terracotta army fragment classification method based on deep learning, and the method comprises the steps: firstly carrying out the preprocessing of a terracotta army fragment data set, removing the background of a picture, cutting off the redundant white edges of the picture, unifying the specification images of the picture, and carrying out the normalization processing;then taking all the processed samples as a training set, randomly selecting part of the samples for labeling, importing the samples into the improved convolutional sparse auto-encoder network in batches for unsupervised iterative training until loss is small enough and tends to be stable, and obtaining output of a bottleneck layer as an effective feature; constructing an anchor graph according tothe effective features, and learning an anchor graph model by using an anchor graph semi-supervised classification algorithm based on a density peak value to obtain a prediction value of an unmarkedsample; and finally, selecting the category with the highest confidence as a prediction label, and comparing the prediction label with a real label to calculate classification accuracy. The method isconvenient to operate, and the problem that an existing method is not high in classification accuracy is solved.

Description

technical field [0001] The invention belongs to the field of cultural relics protection and the like, and relates to deep learning technology, in particular to a method for classifying fragments of Terracotta Warriors based on deep learning. Background technique [0002] In the field of cultural relic excavation and protection, the precise classification of cultural relic fragments is an important research content in this field. The terracotta warriors are the funeral objects of Qin Shihuang. They are huge in number, different in shape, and full of touching artistic charm. They are a golden business card of the splendid civilization of ancient my country. The Terracotta Warriors are rigid bodies of pottery, which are easily broken and broken if they are not well preserved. After natural wind erosion and historical changes, there are very few Qin Warriors pottery that can be completely preserved. The classification of the fragments of the Terracotta Warriors is an important ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/2136G06F18/241G06F18/214
Inventor 王小凤邓胡承张鹏飞刘雨萌刘喆杨稳
Owner NORTHWEST UNIV(CN)
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products