Enhanced Criminisi Algorithm with P-Laplace Operator and PSNR Evaluation

The enhanced Criminisi algorithm optimizes the original data term calculation by introducing the P-Laplace operator, significantly improving image inpainting accuracy. In traditional Criminisi algorithms, gradient information primarily determines the priority term during patch selection, whereas the improved version employs the P-Laplace operator to better capture local structural features of images—particularly excelling in texture and edge regions. From an implementation perspective, this involves calculating anisotropic diffusion patterns through iterative PDE solutions, where the parameter p controls the diffusion intensity across different image structures.

Furthermore, the enhanced algorithm incorporates PSNR (Peak Signal-to-Noise Ratio) computation functionality for objective quality assessment of inpainting results. The PSNR module quantitatively compares differences between original and inpainted images by calculating mean squared error (MSE) and converting it to a logarithmic scale, providing measurable metrics for algorithm optimization. This feature is particularly valuable in high-precision applications such as cultural heritage digitization or medical image processing, where developers can implement automated quality validation loops using threshold-based PSNR checks.

Overall, these improvements maintain the algorithm's computational efficiency while enhancing visual coherence and detail restoration capabilities, making it particularly advantageous for handling complex image缺损 scenarios. The implementation typically involves parallel processing of priority maps and optimized data structures for patch matching, ensuring scalable performance for high-resolution images.