Multi-modal data quality evaluation method based on deep learning
By using deep learning methods to extract semantic features and perform cross-modal joint encoding of multimodal data, the limitations of traditional methods in multimodal data evaluation are overcome, and highly accurate multimodal data quality assessment is achieved.
Patent Information
- Application Number
- CN202511299320.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-12
- Publication Date
- 2025-10-17
- Estimated Expiration
- 2045-09-12
AI Technical Summary
Traditional database data quality assessment methods have limitations when facing multimodal data. They are unable to handle the complex relationships between multimodal data, lack fine-grained hash cross-modal encoding, have difficulty adapting to dynamically changing data, are highly subjective, and lack standardization and automation.
A deep learning-based method is used to obtain the first modality recording data and the second modality recording data marked as aligned through the acquisition engine, and semantic feature extraction and cross-modal joint encoding are performed. The fine-grained hash algorithm and bidirectional reconstruction offset evaluation are used to generate multimodal data quality assessment results.
It improves the accuracy of multimodal data quality assessment, implements an end-to-end deep learning framework, addresses the shortcomings of traditional methods in multimodal data assessment, and enhances the standardization and automation capabilities of assessment.