Multi-functional image Stylization and portrait dispose based on traditional image dispose and deep learning

Abstract

This paper studies the multi-function image based on traditional image processing and deep learning stylized processing method with the portrait, aims to develop the most efficient and convenient image processing system. The content of the research includes two core modules: image stylization and portrait dispose, which combines traditional image dispose techniques with deep learning method. Image stylization module adopts VGG19 network to realize image style transfer, and combines OpenCV technology to realize image cartoonization dispose. Portrait processing module through deep learning algorithms for detecting like cutting and exposed skin. System USES Qt GUI interface technology integration, and through the multithreaded processing efficiency. Results show that a system in the processing efficiency and quality of output achieves the expected effect, provides users with efficient and diversified image processing solutions.