GPU-ACCELERATED IMAGE RETEXTURING IN GRADIENT DOMAIN
Ping Li, Hanqiu Sun, Jianbing Shen · 2010
This paper presents the novel GPU-accelated image retexturing approach for both high and low dynamic range images using our newly invented fast NLM filtering. Integrating the fast Maclaurin polynomial kernel filter and the latest GPU-CUDA acceleration, our approach is able to produce real-time high quality retexturing for objects of the interest, while preserving the original shading and similar texture distortion. We apply our revised NLM filtering to the initial depth map to ensure smoothed depth field for retexturing. Our approach using GPU-based fast NLM filtering is designed in parallel, and easy to develop on latest GPUs. Our testing results have shown the efficiency and satisfactory performance using our approach.