View source for 18450463. METHOD, COMPUTER DEVICE, AND STORAGE MEDIUM, FOR FEATURE FUSION MODEL TRAINING AND SAMPLE RETRIEVAL simplified abstract (Tencent Technology (Shenzhen) Company Limited)

=METHOD, COMPUTER DEVICE, AND STORAGE MEDIUM, FOR FEATURE FUSION MODEL TRAINING AND SAMPLE RETRIEVAL=


==Organization Name==

[[:Category:Tencent Technology (Shenzhen) Company Limited|Tencent Technology (Shenzhen) Company Limited]]


[[Category:Tencent Technology (Shenzhen) Company Limited]]

==Inventor(s)==

[[:Category:Hui Guo of Shenzhen (CN)|Hui Guo of Shenzhen (CN)]][[Category:Hui Guo of Shenzhen (CN)]]

==METHOD, COMPUTER DEVICE, AND STORAGE MEDIUM, FOR FEATURE FUSION MODEL TRAINING AND SAMPLE RETRIEVAL - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18450463 titled 'METHOD, COMPUTER DEVICE, AND STORAGE MEDIUM, FOR FEATURE FUSION MODEL TRAINING AND SAMPLE RETRIEVAL

==Simplified Explanation==

The patent application describes a method for training a feature fusion model and retrieving samples. Here are the key points:

* The method starts by inputting a training sample into an initial feature fusion model to obtain a training semantic feature and a training global feature.
* Classification and recognition are performed based on the training semantic feature to obtain an initial training category.
* The training semantic feature and the training global feature are spliced together to obtain a spliced training feature.
* Autocorrelation feature calculation is performed on the spliced training feature to obtain an autocorrelation feature.
* Self-attention weight calculation is performed based on the autocorrelation feature to obtain a self-attention weight.
* The spliced training feature is adjusted using the self-attention weight to obtain a fused training feature.
* The initial feature fusion model is updated based on the training global feature, training semantic feature, fused training feature, initial training category, and training sample category label.
* The process iterates in a loop to obtain a target fusion model.

Potential applications of this technology:

* Image recognition and classification systems
* Natural language processing and sentiment analysis
* Speech recognition and transcription
* Video analysis and object detection

Problems solved by this technology:

* Improved accuracy and performance in feature fusion models
* Efficient training and retrieval of samples
* Enhanced classification and recognition capabilities

Benefits of this technology:

* Higher accuracy in classification and recognition tasks
* Improved efficiency in training and retrieval processes
* Enhanced feature fusion capabilities for complex data analysis


==Original Abstract Submitted==

A method for feature fusion model training and sample retrieval includes: inputting training sample into an initial feature fusion model to obtain a training semantic feature and a training global feature, performing classification and recognition based on the training semantic feature to obtain an initial training category, splicing the training semantic feature and the training global feature to obtain a spliced training feature, performing autocorrelation feature calculation based on the spliced training feature to obtain an autocorrelation feature, performing self-attention weight calculation based on the autocorrelation feature to obtain a self-attention weight, and adjusting the spliced training feature through the self-attention weight to obtain a fused training feature; and updating the initial feature fusion model based on the training global feature, the training semantic feature, the fused training feature, the initial training category, and a training sample category label, and performing a loop iteration to obtain a target fusion model.



[[Category:G06V10/80]]
[[Category:G06V10/764]]
[[Category:G06V10/774]]