Active vision enhancement of new media images based on semi supervised feature fusion algorithm

• Published in: Signal, image and video processing Vol. 18; no. 8-9; pp. 6221 - 6237
• Main Author: Liu, Dan
• Format: Journal Article
• Language: English
• Published: London Springer London 01.09.2024; Springer Nature B.V
• Subjects: Algorithms; Color; Computer Imaging; Computer Science; Feature extraction; Image enhancement; Image Processing and Computer Vision; Image quality; Machine learning; Multimedia Information Systems; Original Paper; Pattern Recognition and Graphics; Regularization; Semi-supervised learning; Signal,Image and Speech Processing; Vision; Visual effects; Semi supervised learning method; Multiple graphical model; HSV model; Regularization framework; RBG color space
• ISSN: 1863-1703; 1863-1711
• DOI: 10.1007/s11760-024-03309-8

In order to improve the image visual effect, a new media image active vision enhancement method based on semi supervised feature fusion algorithm is proposed. Firstly, the image feature extraction is carried out from the three perspectives of color feature, texture feature and image subject feature. The image feature is projected into the single image model, the regularization framework is established, and the multiple graphical model based on semi supervised learning method is constructed to complete the image feature fusion and determine the image enhancement strength. HSV model is introduced to decompose the image into three channel components of H, S and V. Through the adaptive component adjustment of the three channels and RBG color space conversion, the active visual enhancement of the image is achieved. The test results show that the proposed method effectively avoids the overexposure and local distortion of the image, and retains the clarity of image details to the greatest extent. The overall quality of the enhanced image is high.