CAJ | 학술논문

图像质量评价是图像和视频处理系统中的基本算法. 通过对人眼视觉特性及图像的边缘特征的分析研究,认为梯度矢量更能反映图像边缘的本质特性,结合人眼视觉的多尺度特性及视觉感兴趣区域的加权,提出了一种基于梯度矢量的压缩图像质量评估方法. 仿真实验表明,该评估方法符合人眼的视觉特征,与主观评价结果具有更好的一致性,可广泛应用于JPEG和JPEG2000压缩图像质量评估.
도상질량평개시도상화시빈처리계통중적기본산법. 통과대인안시각특성급도상적변연특정적분석연구,인위제도시량경능반영도상변연적본질특성,결합인안시각적다척도특성급시각감흥취구역적가권,제출료일충기우제도시량적압축도상질량평고방법. 방진실험표명,해평고방법부합인안적시각특정,여주관평개결과구유경호적일치성,가엄범응용우JPEG화JPEG2000압축도상질량평고.
Image quality assessment is fundamental importance for image or video processing system . Through the research and analysis human visual system features and the edge feature of image , this paper consider that the gradient vector better reflect the essential characteristics of image edge feature .Com-bined with the multi -scale feature and weighted interest region of human visual system , a new com-pressed image quality assessment method based on gradient vector is proposed .The proposed method is simple in calculation .Simulation results upon image databases show that the evaluation method is in ac-cordance with characteristics of human visual system and consistency with subjective evaluation results better.In addition the method can be widely used to assess the quality of JPEG and JPEG 2000 com-pressed images .