Multimodal medical image fusion using convolutional neural network and extreme learning machine

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computer Vision

Background:

• Multimodal medical imaging improves diagnosis but faces limitations with individual modalities.
• Image fusion offers a solution by combining complementary information from different imaging techniques.

Purpose of the Study:

• To propose a novel fusion method for multimodal medical images.
• To address the limitations of traditional Convolutional Neural Networks (CNNs) like high computational costs and human intervention.

Main Methods:

• A novel fusion method combining CNNs and Extreme Learning Machines (ELMs) is developed.
• A Convolutional Extreme Learning Machine (CELM) model is constructed by integrating ELM into CNN.
• CELM is used to extract and capture image features from multiple angles for fusion.

Main Results:

• The proposed CELM method effectively extracts and fuses significant features from multimodal medical images.
• Experimental results demonstrate enhanced accuracy in lesion detection and localization.
• The method shows superior performance in both subjective visual assessment and objective criteria.

Conclusions:

• The novel CELM-based fusion method offers an effective solution for multimodal medical image analysis.
• This approach improves diagnostic accuracy and visual quality in medical imaging.
• The proposed method outperforms current state-of-the-art techniques for image fusion.