Cross-modal retrieval method based on adversarial learning and asymmetric hashing
A cross-modal and asymmetric technology, applied in the field of cross-modal retrieval, which can solve the problem that the hash code is not optimal, and the modal data cannot be fully extracted.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0051] In order to better express the cross-modal retrieval method based on adversarial learning and asymmetric hashing proposed in the present invention, in conjunction with the drawings and specific implementations, a 224×224 pixel picture and its corresponding text description are taken as an example. , To further explain the present invention.
[0052] figure 1 It is the overall flow chart of the present invention, including five parts: data preprocessing, initialization model framework, model training, hash code generation of each modal data, and retrieval stage.
[0053] Step 1. Data preprocessing. Divide the cross-modal data set into two parts: training set and test set. Each data instance contains picture-text pairs and their corresponding labels;
[0054] Step 2. Initialize the model framework. figure 2 It is a model framework designed in the present invention. The framework includes a cross-modal feature extraction module, an attention module guided by a counter-network,...
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.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap