Research on Image Quality Enhancement Technology of New Media Based on Computer Vision

In response to the significant impact of image quality on user experience and information dissemination efficiency in the new media environment, this study proposes an image quality enhancement scheme using an Adaptive Multi-Scale Convolutional Neural Network (AMS-CNN). The research captures detail information at different scales in the image through a multi-scale convolution module and combines it with an adaptive weighting fusion strategy to achieve precise image quality enhancement. Experimental results show that compared to traditional bilateral filtering, the super-resolution network SRGAN, and another multi-scale convolutional network MS-CNN, AMS-CNN has achieved significant advantages in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM), with values of 32.150 dB and 0.912 respectively. User satisfaction surveys also indicate that AMS-CNN received the highest average satisfaction score of 4.6. This study not only provides effective technical support for new media image quality enhancement but also promotes the further application and development of computer vision technology in the field of image processing.