Jigsaw puzzle generation method and device

The invention discloses a jigsaw puzzle generation method. The method comprises the following steps: acquiring a size proportion of an original image to be spliced and a generated image; obtaining a maximum saliency region of the original image by using a saliency detection method; generating candidate cutting images according to the saliency region and a preset cutting method; performing aesthetic scoring on the candidate cut images, and selecting the cut image with the highest score; performing face detection on the cut image with the highest score to obtain a display weight; and filling the cut image into a splicing template according to the significant weight. Splicing effect conforming to user preferences can be more effectively and automatically generated, and user experience is improved. In addition, on the other hand, the invention provides a jigsaw puzzle generation method device. 本发明公开了一种拼图生成方法，包括：获取需要拼图的原始图像与生成图像的尺寸比例；使用显著性检测方法，获取原始图像的最大显著性区域；根据显著性区域与预设的裁剪方法生成候选裁剪图像；将候选裁剪图像进行美学评分，并选取评分最高的裁剪图像；将评分最高