Micro-video classification method and system based on missing data completion
A technology of missing data and classification methods, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as reducing the accuracy of micro-video classification tasks, achieve strong semantic representation capabilities, and ensure accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0059] Such as figure 1 As shown, this embodiment provides a micro-video classification method based on missing data completion. This embodiment uses this method as an example to illustrate the application of the server. It can be understood that this method can also be applied to terminals, and can also be applied to It includes terminals, servers and systems, and is realized through the interaction between terminals and servers. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or it can provide cloud services, cloud database, cloud computing, cloud function, cloud storage, network server, cloud communication, intermediate Cloud servers for basic cloud computing services such as software services, domain name services, security service CDN, and big data and artificial intelligence platforms. The terminal may be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart spe...
Embodiment approach
[0064] As one or more implementations, the two-way cyclic generative adversarial network includes: three cyclic generative adversarial networks, each cyclic generative adversarial network includes two directions, generating the second modality from the first modality and generating the second modality from the second modality Modality generates a first modal, where the first modal is a visual or audible modal or a text modal, the second modal is a visual or audible modal or a textual modal, and the first and The second modality is not the same.
[0065] Specifically, such as image 3 Shown: The features of the three modalities of the micro-video pass through three sets of cyclic generative adversarial networks to generate feature representations of other modalities respectively. Each set of recurrent generative adversarial networks includes two directions, generating modal B from modal A and generating modal A from modal B (generating modal C from modal B and generating modal...
Embodiment 2
[0092] This embodiment provides a micro-video classification system based on missing data completion.
[0093] A micro-video classification system based on missing data completion, including:
[0094] The classification module is configured to: based on the micro-videos with partial modal data missing, adopt the trained micro-video classification network to obtain the classification result of the micro-videos with partial modal data missing;
[0095] Model building module, it is configured as: described micro-video classification network comprises: Based on the micro-video that there is partial modal data missing, adopts two-way circulation to generate confrontation network, obtains the missing modality that complements micro-video; The original micro-video The missing modals of the modal and complementary micro-videos pass through the common subspace learning module to extract the semantic feature representation vector of the visual modality, the semantic feature representati...
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