Key points are not available for this paper at this time.
To realize a high-quality, real-time naked-eye 3D light-field display, this paper proposes a perceptual optimization method based on multi-plane reconstruction and visual component modulation to address the discontinuous depth geometric representation and insufficient interlayer perceptual consistency issues in existing light-field reconstruction methods. The depth optimization of image details and color levels is achieved by introducing a depth-image joint-guided layering strategy in the multiplanar reconstruction stage to enhance depth continuity, and by collaboratively modulating the structure and color of the visual components. The experimental results show that this method outperforms the traditional algorithm in SSIM, PSNR, and LPIPS metrics, effectively improving color consistency and overall visual comfort, and providing key technical support for the future generation and display of high-quality light-field images.
Qiu et al. (Fri,) studied this question.