Medical image fusion plays a central role by integrating information from multiple sources into a single, more understandable output. We propose a real-time image fusion method using pre-trained neural networks to generate a single image containing features from multi-modal sources. Our method can be applied to any number of input sources.