18456312. PERCEPTION NETWORK AND DATA PROCESSING METHOD simplified abstract (HUAWEI TECHNOLOGIES CO., LTD.)
Contents
PERCEPTION NETWORK AND DATA PROCESSING METHOD
Organization Name
Inventor(s)
PERCEPTION NETWORK AND DATA PROCESSING METHOD - A simplified explanation of the abstract
This abstract first appeared for US patent application 18456312 titled 'PERCEPTION NETWORK AND DATA PROCESSING METHOD
Simplified Explanation
The abstract describes a perception network for artificial intelligence applications, specifically in the field of feature extraction. The network includes multiple blocks that perform convolution processing on input data to obtain target feature maps. These feature maps are then further processed and concatenated to create a final concatenated feature map.
- The perception network is designed for artificial intelligence applications.
- It includes a feature extraction network that performs convolution processing on input data.
- The first block in the network obtains M target feature maps.
- At least one second block performs convolution processing on M1 target feature maps to obtain M1 first feature maps.
- A target operation is used to process M2 target feature maps to obtain M2 second feature maps.
- The feature extraction network then concatenates the M1 first feature maps and M2 second feature maps to obtain a concatenated feature map.
Potential Applications
- This technology can be used in various artificial intelligence applications that require feature extraction, such as image recognition, natural language processing, and speech recognition.
- It can be applied in autonomous vehicles for object detection and tracking.
- The perception network can be used in robotics for object recognition and manipulation.
Problems Solved
- The perception network solves the problem of efficiently extracting meaningful features from input data in artificial intelligence applications.
- It addresses the challenge of processing and combining multiple target feature maps to create a comprehensive feature representation.
Benefits
- The feature extraction network allows for efficient and effective extraction of features from input data.
- By concatenating the first and second feature maps, the network creates a more comprehensive and informative feature representation.
- The perception network can improve the accuracy and performance of various artificial intelligence tasks by providing better feature extraction capabilities.
Original Abstract Submitted
This disclosure discloses a perception network. The perception network may be applied to the artificial intelligence field, and includes a feature extraction network. A first block in the feature extraction network is configured to perform convolution processing on input data, to obtain M target feature maps; at least one second block in the feature extraction network is configured to perform convolution processing on M1 target feature maps in the M target feature maps, to obtain M1 first feature maps; a target operation in the feature extraction network is used to process M2 target feature maps in the M target feature maps, to obtain M2 second feature maps; and a concatenation operation in the feature extraction network is used to concatenate the M1 first feature maps and the M2 second feature maps, to obtain a concatenated feature map.
