This paper focuses on pulse coupled neural network (PCNN) and digital image fusion. Aiming at the existing problems, this paper proposes a real-time deep learning model with dual-channel PCNN fusion algorithm based on image definition. It will also be helpful to digital image forensics. With the integration of the orthogonal color space that conforms to HVS, this algorithm simplifies the traditional PCNN model to a parallel dual-channel adaptive PCNN structure. Also, it can realize the adaptive processing by defining the image definition to be β, the coupled linking coefficient. As the dynamic threshold can be increased exponentially with this method, it can effectively solve the problems. The experimental result proves that our algorithm outperforms the traditional fusion algorithms according to the subjective visual effect or the objective assessment standard.

中文翻译：

---

基于图像定义的深度学习模型在实时数字图像融合中的应用

本文重点研究脉冲耦合神经网络（PCNN）和数字图像融合。针对目前存在的问题，提出了一种基于图像定义的双通道PCNN融合算法的实时深度学习模型。这也将有助于数字图像取证。通过整合符合HVS的正交色彩空间，该算法将传统PCNN模型简化为并行双通道自适应PCNN结构。另外，通过将图像定义为β，可以实现自适应处理。，耦合链接系数。由于该方法可以动态增加动态阈值，因此可以有效地解决这些问题。实验结果证明，根据主观视觉效果或客观评估标准，该算法优于传统融合算法。