Image Inpainting via Stochastic Dynamics
2026-07-27 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors present a new way to fill in missing parts of images without using any trained neural networks. Their method uses data from reference images to guide a step-by-step process that gradually reconstructs the missing pixels. They tested this approach on several datasets and found it gave better or comparable results to traditional fill-in methods, especially preserving important structures. This shows that using actual data patterns directly can be effective for image restoration tasks.
image inpaintingstochastic differential equationnon-parametric methodreverse-time diffusionkernel-weighted correctionreference datasetPSNRSSIMMNISTCelebA
Authors
Jiaqi Kuang, Zihao Guo, Zhongmin Qian
Abstract
Image inpainting aims to recover missing regions while preserving structural consistency. We propose a non-parametric method without network training based on data-guided stochastic dynamics. Starting from a masked image, the missing pixels are evolved through a reverse-time stochastic differential equation with a kernel-weighted correction estimated directly from a reference dataset. This empirical correction guides the reconstruction toward high-density regions of the data distribution without training a neural network or fitting a parametric density model. Experiments on MNIST, Fashion-MNIST, and MVTec show that the proposed method outperforms Mean Fill, Telea, and Navier-Stokes inpainting in PSNR, SSIM, and visual quality. On CelebA, it remains competitive and produces plausible completions for structure-sensitive occlusions. These results demonstrate the effectiveness of empirical reference statistics as a non-parametric prior for image inpainting.