MMICap

We presents MMICap, a deep-learning-based solution that uses multimodal inputs comprised of scientific images paired with detailed texts to automatically generate captions that effectively summarize the content depicted in these images.