基于小波变换的自适应脉冲耦合神经网络图像融合

针对同一场景多聚焦图像的融合问题，提出了一种基于小波变换的自适应脉冲耦合神经网络（PCNN）图像融合方法。首先，对源图像进行小波分解，得到不同尺度下的子带图像；然后，在小波域中利用PCNN的同步脉冲激发特性，制定基于PCNN的融合规则；使用不同尺度下的小波系数的拉普拉斯能量（EOL）作为对应神经元的链接强度，经过PCNN点火得到源图像在小波域中的点火映射图；通过判决选择算子，选择点火次数多的小波系数作为对应的融合系数，然后进行区域一致性检验，获到最终的融合系数；最后，对融合后的系数进行小波逆变换得到融合图像。实验结果表明，该方法更有效地提取原始图像的特征信息，提高融合图像的视觉效果，在主观视觉效果与客观性能指标上均优于传统的图像融合方法。

Image fusion based on wavelet transform and adaptive PCNN

Concerning the fusion problem of multi-focus image with the same scene, an algorithm of image fusion based on wavelet transform and adaptive Pulse Coupled Neural Network (PCNN) was proposed. Firstly, original images were decomposed by wavelet transform, and the sub-band images at different scales were obtained. Secondly, a fusion rule was given through making use of synchronous pulse bursts. This method used Energy of Laplacian (EOL) of wavelet coefficients at different scales as the linking strength of the corresponding neuron. After the processing of PCNN with the adaptive strength, new fire mapping images in wavelet domain were obtained. According to the fire mapping images, the fusion coefficients were decided by the compare-select operator, and then the region consistency test was used on the fusion coefficients to obtain the final fusion coefficients. Finally, fusion images were obtained by wavelet inverse transform. The experimental results illustrate that this algorithm is efficient to extract feature information from the original images and improves fusion images. It outperforms the conventional methods in subjective vision effect and objective performance index.