Multi-angle image quality evaluation method and system

The invention discloses a multi-angle image quality evaluation method and system. The method comprises the steps: acquiring original image data, wherein the original image data comprise at least one piece of facial image data; splitting the original image data to obtain local image data; training an image quality evaluation model according to the local image data and image label data, wherein the image label data comprises an image quality description vector and an image illumination description vector; and performing prediction through the image quality evaluation model to generate an image evaluation result. The technical problem that in the prior art, the image quality cannot be evaluated from the aspects of local parts, light rays and the like is solved. 本发明公开了一种多角度图像质量评价方法及系统。其中，该方法包括：获取原始图像数据，其中，所述原始图像数据包括至少一张面部图像数据；拆分所述原始图像数据，得到局部图像数据；根据所述局部图像数据和图像标签数据训练图像质量评价模型，其中，所述图像标签数据包括：图像质量描述向量、图像光照描述向量；通过所述图像质量评价模型进行预测，生成图像评价结果。本发明解决了现有技术中图像质量无法从局部、光线等角度进行评价的技术问题。