Find in Library

Aiming at the problem that a no-reference (NR) image quality assessment algorithm is not accurate enough to predict the image quality at present, we proposed a method of NR image quality assessment based on spatial domain coding (SDC). We extracted the spatial structure features of the image on different bit planes. The extracted features quantify the structural information between pixels, which can more accurately reflect the distortion degree of the image. We used a neural network to establish an image quality assessment model. The experimental results show that the proposed image assessment algorithm is more accurate than the existing mainstream image quality assessment algorithms and highly consistent with subjective perception of human eyes. Finally, the proposed algorithm is applied to the auto focus of the camera verifying the practicability and accuracy of the algorithm.