Chinese paper cutting identification method based on space constraint characteristic selection and combination thereof
A feature selection, space constraint technology, applied in the field of image recognition
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach
[0061] After the classification error of each component of the candidate feature V is obtained, it is sorted according to the classification error, and the first Z components with the smallest error in V are selected to form a distinctive feature of the j-th type of paper-cut image. The specific implementation is as follows:
[0062] Input: training sample set S={s k |1≤k≤M}; annotation matrix T={t ij ∈{0, 1}|1≤i≤M, 1≤j≤N}; the feature set to be selected V={v ik |1≤i≤M, 1≤k≤L}
[0063] Output: Z most discriminative components in feature set V
[0064] step:
[0065] 1. Initialize weight W={w ij |1≤i≤M, 1≤j≤N}
[0066] 2. Calculate the arithmetic mean of the sample candidate features C={c k |1≤k≤L}
[0067] 3. Repeat the following operations from k=1 to L:
[0068] a) Calculate:
[0069] FSC ik = 1 if ( sor...
Embodiment 1
[0077] 1) Constructing a paper-cut dataset: 246 Chinese paper-cut images were collected from the Internet. According to these paper-cut art themes, such as figure 2 Shown divides the dataset into four categories: animals, portraits, people, and text. Use 80% of the samples as the training set and 20% as the test set;
[0078] 2) To train the training set samples, the specific steps are as follows:
[0079] a) Preprocess the training set samples, extract SIFT features, and use the K-Means clustering algorithm to extract visual words to form the initial features of the samples;
[0080] b) The method of combining spatial pyramid matching and context-dependent histograms is used to process the initial features of the samples to form candidate features of the samples;
[0081] c) Using AdaBoost-based feature selection and combination technology to process sample candidate features to obtain sample distinguishing features;
[0082] d) Characterize the class by the class center...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com