摘要
研究图像优化特征提取问题,针对自然场景图像由于在拍摄过程中造成图像模糊不清晰等缺陷,基于神经网络提出了一种新的综合多种特征和判别标准的图像蜕化算法。首先对特定的蜕化成因,确定不同的特征提取方法;然后针对不同特征值采用不同的判定标准对图像模糊的程度进行初步判断,最终通过神经网络学习算法,将不同特征及其判定标准融合为最终的图像模糊判定准则。仿真结果显示所得到的模糊估计值与图像真实模糊类型与参数保持一致,有效地提高了图像蜕化的质量,并且能准确地判断出图像模糊的程度和可能的蜕化种类,为图像恢复或者后续图像处理的决策提供有利信息。
In this paper,we addressed the problem of no-reference blur assessment for digital pictures.We started with the generation of a large real image database containing pictures taken by human users in a variety of situations,and the conduction of subjective tests to generate the ground truth associated to those images.Based on this ground truth,we selected a number of high quality pictures and artificially degraded them with different intensities of simulated.We proposed a paradigm for blur evaluation in which an effective method was pursued by combining several metrics and low-level image features.We tested this paradigm by designing a no-reference quality assessment algorithm for blurred images which combined different metrics in a classifier based on a neural network.Experimental results show that this leads to an improved performance that reflects the images＇ ground truth better.
出处
《计算机仿真》
CSCD
北大核心
2011年第12期253-255,263,共4页
Computer Simulation
关键词
图像模糊
神经网络
评估
蜕化
Image blur
Neural network
Assessment
Degenerate
