基于机器学习的素描图像处理技术

图像素描化在激光雕刻领域中发挥着极其重要的作用,传统的图像素描处理方法是不同的图像采用同一种算法进行处理,存在适用率不高和部分参数不通用问题。文章提出一种基于机器学习的素描图像处理技术,通过搭建素描图像处理神经网络系统,使用构建的素描图片数据库对模型进行多次迭代训练,最后达到预期素描处理效果,提高了图片素描化处理的普适性。

Image sketching plays an extremely important role in the field of laser engraving. The traditional image sketch processing method is that different images are processed by the same algorithm, which has the problems of low application rate and incompatibility ofsome parameters. In this paper, an image sketch processing technology based on machine learning is proposed. By building a sketch image processing neural network system, the sketch image database is used to give the model the iterativetraining for many times, and finally the expected sketch processing effect is achieved. The universality of image sketch processing is improved.