Image lighting method and device

Abstract

Embodiments of the present invention provide an image lighting method and device, the method comprising: determining the convolutional neural network corresponding to the lighting operation type of the initial image; and obtaining local feature information and global feature information of the initial image through the convolutional neural network ; Through the convolutional neural network, based on local feature information and global feature information, the fusion feature information of the initial image is obtained; through the convolutional neural network, based on the brightness component map of the initial image, the maximum pooling result map and the minimum pooling result map are obtained ;Based on the fusion feature information, the maximum pooling result map and the minimum pooling result map, the bilateral grid matrix of the initial image is obtained; according to the bilateral grid matrix, the initial image is affine transformed to obtain the target image, and the target image is the original image according to Lighting operation type Lighted image. The image lighting method provided by the embodiment of the present invention has a better effect.

Description

technical field [0001] The invention relates to image processing technology, in particular to an image lighting method and device. Background technique [0002] As the camera pixels of terminal devices (such as mobile phones, tablets, bracelets, etc.) continue to improve, more and more users frequently use terminal devices to take portraits and to beautify images after portraits to obtain ideal portrait effects. Ideal portrait shooting requires professional lighting, applying proper amount of light to the face to achieve a better photo effect, but it is difficult to achieve the ideal lighting environment in most actual shooting scenes, so the portraits in the actual shooting images There are usually problems such as dark face, lack of three-dimensional effect, and uneven facial lighting. [0003] To solve the above problems, a common solution is to perform 3D modeling of the portrait based on the portrait photo, adjust the lighting on the portrait, render the image, and obt...

AI Technical Summary

Problems solved by technology

[0004] However, portrait modeling is greatly affected by face recognition, and when face recognition is inaccurate, the modeling error is also large

Therefore, the existing image lighting scheme is complex and inefficient

Images