基于神经网络的图像融合效果综合评价

针对单因素指标评价图像融合效果时只考虑融合图像某一方面的特征，缺乏全局性等问题，提出了一种基于神经网络的图像融合效果综合评价方法。该方法采用图像熵值、空间频率值、交叉熵平均值、均方差平均值构成单因素评价指标集，利用模糊积分求出单因素指标融合效果评价值，并以此作为神经网络的输入对其权值进行训练，最终获得综合评价指标。通过大量红外与可见光图像融合的效果评价实验，证明该方法评价结果合理，主客观评价有较好的一致性。

Comprehensive Evaluation of Image Fusion Based on Artificial Neural Network

Aiming at the problem of single-factor evaluation on the effort of image fusion, which only consider some feature of the fusion image and lack the global characteristic, a comprehensive evaluation method based on artificial neural networks is proposed. Entropy, spatial frequency, cross entropy and mean-square error of the image is selected to form a set of single-factor evaluation index. First, get the evaluation value of single-factor index through fuzzy integral, and then take them as inputs of artificial neural network to train their weights. Thus the comprehensive evaluation index is achieved. The method is applied to series of experiments on the evaluation of infrared and visible light image fusion, which prove that the evaluation result is reasonable and consistent with subjective evaluation.