【摘要】 在视频图像质量客观评价方法中,基于结构相似度(SSIM)模型算法由于具有对像素偏移不敏感、计算简单等特点,应用较为广泛。但SSIM模型的缺点是不能较好地评价严重模糊的受损图像。因此,提出了一种基于边缘结构相似度(ESSIM)的视频图像质量评价方法,将边缘梯度提取作为图像的主要结构信息。实验结果表明,ESSIM模型比SSIM模型更符合人眼视觉系统特性,能更好地反映图像的主观质量。更多还原

【Abstract】 In the objective methods of video image quality,structural similarity(SSIM) model algorithm is widely used based on that it isn’t sensitive to the pixel offset and calculates easily.But the SSIM model cannot evaluate the blurred image well which is damaged severely.In this paper,an evaluation method based on edge similarity structure of the video quality(ESSIM) is proposed.In the method,the gradient edge detection is as the main structure of the image information.The results show that ESSIM model is more coincidence on human visual system characteristics than the SSIM model,and reflects the subjective video quality better.更多还原