The data volume of high-definition (HD) photographic images is proliferating, which puts forward higher requirements for efficient compression and high-quality recovery techniques, for which an innovative technique for all-media converged HD image compression and recovery based on improved multiple-external optimize algorithm is proposed. By improving the traditional genetic algorithm to acquire more advanced computational precision and faster convergence speed, the improved arithmetic is applied to optimizing the weightings and value range of the BP neuro mesh to enhance the generalization perform of the neuro mesh. And the combined pattern is applied to the HD camera image compression and recovery under the all-media convergence, which calculates and retains the image compression and recovery data, and reconstructs the HD image from the data blocks. The average GD and IGD of this paper’s method are 0.0312 and 0.0703 for image compression and recovery strategy optimization, which are more advanced than the comparison ways. With the improvement of the degree of graphic compression, the image compression evaluation indexes of the method are improved to different degrees. For image 1, the PSNR of image compression of this method is 30.45db and MS-SSIM is 0.91, which is better than the pass comparison method. In five types of image recovery tasks, the PSNR of this paper’s model is between 31.69 and 34.76db, and the SSIM is between 0.846 and 0.924, and the image recovery tasks can be accomplished excellently.
