No-reference image quality evaluation method based on causal distillation

The invention provides a no-reference image quality evaluation method based on causal distillation. The method comprises the following implementation steps: acquiring a training sample set and a test sample set; constructing a no-reference image quality evaluation network model based on causal distillation; iteratively training the no-reference image quality evaluation network model based on causal distillation; and obtaining a no-reference image quality evaluation result. According to the causal distillation sub-network constructed according to the causal distillation strategy, the feature weight is adjusted in the training process, the problem of disastrous forgetting caused by distillation loss in transfer learning in the prior art is effectively avoided, the model has good prediction precision when facing a plurality of data sets, and the prediction efficiency is improved. And the precision of non-reference image quality evaluation is effectively improved. 本发明提出一种基于因果蒸馏的无参考图像质量评价方法，实现步骤为：获取训练样本集和测试样本集；构建基