图像处理的方法、装置、电子设备及存储介质
By acquiring information about the positions of nearby lights in the dome's light field and using a neural network model to process the image, the problems of highlights and ghosting caused by linear interpolation in the reconstruction of dome light field images were solved, achieving efficient image rendering effects.
CN116309111BActive Publication Date: 2026-07-17SHANGHAI TECH UNIV +1
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANGHAI TECH UNIV
- Filing Date
- 2023-01-10
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In existing technologies, due to the sparse placement of lights during the reconstruction of dome light field images, only linear interpolation methods can often be used, resulting in incorrect highlights and ghosting, which affects the rendering effect.
Method used
By acquiring the location information and images of neighboring lights at the target light location, and using the encoder and decoder of a pre-built neural network model for image processing, the image corresponding to the target light location is obtained, avoiding linear interpolation.
Benefits of technology
Acquire high-quality images of target light locations in a short time to avoid incorrect highlights and ghosting, thereby improving rendering results.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN116309111B_ABST
Abstract
本申请提供了一种图像处理的方法、装置、电子设备及存储介质，其中，在图像处理的方法中，首先获取目标灯光位置的邻近灯光位置的位置信息和邻近灯光位置对应的图像。然后调用预先构建的图像处理模型的编码器对图像进行处理，得到图像的图像特征。最后调用图像处理模型的解码器对图像特征和位置信息进行处理，得到目标灯光位置对应的图像。由此可知，利用本申请的方法，通过预先构建的图像处理模型对邻近灯光位置的位置信息和邻近灯光位置对应的图像进行处理，利用神经网络进行图像数据增强，可以在较短的时间内得到目标灯光位置对应的图像，不存在不正确高光与伪影问题。
Need to check novelty before this filing date? Find Prior Art