特征处理方法、特征提取方法、装置及电子设备
By determining the channel influence coefficients of the data feature images, arranging and splitting the channel set in order, and using a convolutional structure that adapts to the receptive field, the problem of low accuracy in data feature extraction is solved, and richer and more accurate feature representation is achieved.
CN115937528BActive Publication Date: 2026-07-17SHENZHEN XUMI YUNTU SPACE TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENZHEN XUMI YUNTU SPACE TECH CO LTD
- Filing Date
- 2022-11-23
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing technologies have low accuracy in data feature extraction, leading to the loss of important information.
Method used
By determining the influence coefficient of each channel, arranging and splitting the channel set in order, and processing the channel set using convolutional structures with different receptive fields, rich and important information can be obtained.
Benefits of technology
It improves the accuracy of feature extraction and the ability to represent features, and enhances the richness and importance of information.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN115937528B_ABST
Abstract
本公开涉及计算机技术领域，提供了特征处理方法、特征提取方法、装置及电子设备，该特征处理方法包括：确定第一特征图像的每个通道分别对应的影响系数；基于影响系数，对第一特征图像的通道进行排序处理，得到按序排列的通道；在按序排列的通道中，将排序在不同比例区间的通道进行拆分，得到至少两个通道集合，一个通道集合对应一个比例区间；对至少两个通道集合进行卷积处理，得到每个通道集合对应的第二特征图像；其中，通道集合对应的影响系数越大，对通道集合进行卷积处理的卷积结构的感受野越大；基于每个通道集合对应的第二特征图像，得到第三特征图像。对不同通道集合进行不同的卷积处理，有利于获取到更为丰富有效的特征信息。
Need to check novelty before this filing date? Find Prior Art