摘要
提出一种高效的图像区分方法,专注于区分实拍图像与生成图像。该方法在图像预处理过程中,利用了图像尺度及颜色数等几个高效的判断项,将那些容易区分的生成图像召回,减少了后期垃圾图像与文字图像判断项的判断量,减少了图像筛选运行时间。
This paper proposed an efficient graphic distinction algorithm which focused on distinguishing graphics versus photographs.By utilizing several highly efficient entries such as scale and color,the computer-generated graphics could be recalled so that the postprocessing stage of spam images and text images had less workload which significantly reduced the run-time of image filtering
出处
《计算机应用研究》
CSCD
北大核心
2012年第9期3540-3542,3550,共4页
Application Research of Computers
