{"id":181876,"date":"2026-06-03T14:48:10","date_gmt":"2026-06-03T12:48:12","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novosti-interneta\/open-image-denoise-2-5-0"},"modified":"2026-06-03T14:48:10","modified_gmt":"2026-06-03T12:48:12","slug":"open-image-denoise-2-5-0","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/news\/open-image-denoise-2-5-0","title":{"rendered":"Open Image Denoise 2.5.0","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"


\n
\n <\/a><\/p>\n


\n
\n <\/a><\/p>\n

<\/a><\/noindex><\/p>\n

<\/a><\/p>\n

Release published Open Image Denoise 2.5.0<\/strong> \u2014 an open-source library from Intel for noise suppression in images generated by ray tracing. The project is distributed under the license Apache 2.0<\/strong> and is part of the Intel Rendering Toolkit. The v2.5.0 release was published on GitHub on June 2, 2026<\/strong>.<\/p>\n

<\/p>\n

Open Image Denoise<\/a> is designed to remove Monte Carlo noise characteristic of stochastic rendering methods, including path tracing. The library provides a C\/C++ API and a set of neural network filters that can work with both color buffers and additional buffers such as albedo<\/strong> and normal<\/strong>, to better preserve image details.<\/p>\n

<\/noindex><\/p>\n

The main change in version 2.5.0 is a significant performance optimization and reduction in memory consumption on Intel GPUs with XMX support<\/strong> and on processors with AMX-FP16<\/strong>. The developers did not provide specific acceleration numbers, but highlighted this separately as the first change in the release notes.<\/p>\n

Key changes in Open Image Denoise 2.5.0<\/h4>\n