A Parallel Computation Framework For Generating Painterly Style Arts

With the development of computer graphics, multimedia and artificialintelligence technology, academic community and industrial community arebecoming interested in producing non-photorealistic art works, but thereare still many shortcomings and insufficiency in this area, especiallythe painterly rendering technology. On the one hand, painterly renderingalgorithm calculation efficiency is still very low, because of the largeamount calculation. In today s society, multicore processor became moreand more popular, but the traditional painterly rendering algorithm isnot optimized for multicore processor. On the other hand, the algorithmdon t have a good performance, because of the pure color stroke cannotsimulate the pigment of oil painting very well. In addition, painterlyrendering technology is still in beginning, it needs a good software andfriendly user interface to make people accept and use.To overcome these problems, we do a lot of research for painterlyrendering algorithm, and redesign the framework based on the parallelcomputing. For the poor performance, we use stroke texture to enhance therendering effect. We produce a height map from the texture and then usingPhong lighting model to simulate the pigment of oil painting. Moreover,we have designed software for people who don t have professionalknowledge. By using out software, they can produce an artwork easily.