图像处理方法、装置、电子设备及可读存储介质

By using a pre-trained convolutional neural network model to calculate the placement area of ​​sticker materials, the problem of high cost and long time consumption in training key point models in existing AR texture mapping is solved, achieving efficient image processing and improved user experience.

CN115375880BActive Publication Date: 2026-07-17VIVO MOBILE COMM CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
VIVO MOBILE COMM CO LTD
Filing Date
2022-08-17
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

In existing AR mapping technologies, training keypoint models is costly and time-consuming, resulting in low image processing efficiency.

Method used

By using a pre-trained convolutional neural network model, the placement area of ​​the sticker material is determined by calculating the position transformation parameters through feature extraction and combination of template image and user image, without the need to train a keypoint model.

Benefits of technology

It enables the quick and low-cost determination of sticker material placement areas, reducing image processing time and costs, and improving user experience and work efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115375880B_ABST
    Figure CN115375880B_ABST
Patent Text Reader

Abstract

本申请公开了一种图像处理方法、装置、电子设备及可读存储介质，属于人工智能技术领域。提供的图像处理方法，获取包含第一对象的第一用户图，获取与第一用户图对应的模板图，将第一用户图和模板图输入预先训练好的卷积神经网络模型中，得到卷积神经网络模型输出的第一用户图相对于模板图的位置变换参数，根据位置变换参数，确定选取的贴纸素材在第一用户图中的放置区域，然后将该贴纸素材叠加到第一用户图的放置区域中，从而得到包含上述贴纸素材的第二用户图。
Need to check novelty before this filing date? Find Prior Art