网络视频广告图像非显著目标自适应提取仿真

摘要

An evaluation method combing weighting information quantity with human visual characteristic is put forward.The local degree of distortion in advertising image of network video is measured to compute the gradient vector of image and the similarity distribution of gradient vector.Then,the weight function of visual perception is solved.After multiplying the two together,the evaluation index of communication effect of image in corresponding scale can be obtained.Finally,different weights are assigned to different scales.Thus,the extraction of non-significant target of network advertisement image can be realized.Simulation results show that the proposed method has accurate evaluation and strong self-adaptive degree.%对网络视频广告图像非显著目标的提取,能够有效提高广告传达效率.对网络视频广告图像非显著目标的自适应提取,需要根据图像的梯度信息向量,计算梯度矢量相似度分布,完成图像非显著目标提取.传统方法依据相应提取模型给出量化标准,模拟人类视觉系统感知性能衡量图像传达效果,但忽略了计算图像梯度矢量相似度分布,导致提取精度偏低.提出基于信息量加权与人眼视觉特性相结合的评估方法,对网络视频广告图像局部失真程度进行度量,计算图像的梯度信息向量,计算梯度矢量相似度分布;求解视觉感知权重函数,两者相乘即可得到相应尺度的图像传达效果评估指标.对图像不同的尺度分配不同的权重,使得评估结果更加准确.实验结果表明,能够实现对网络视频广告图像非显著目标的提取,自适应程度较强.