FusionNeXt: A ConvNeXtV2 and ResNet Framework Guided by a Hybrid Loss Function for Combining Anatomical and Functional Medical Imagery

View at publisher Summarise with VIRAT
Authors: Pathan Mujahid, Gunnam Suryanarayana, Bandarupalli Gopichand, Thori Kranthikiran
Cited by: 0
References: 17

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Medical images can give doctors valuable information about how a patient is doing when they look at different types of images together, like MRI scans which show structural details of the brain and PET scans which show functions of the brain. In this paper we introduce a new deep learning framework called FusionNeXt, which will combine MRI and PET scans into one new image (the fused image) that has more useful information than either image would alone. FusionNeXt uses ConvNeXtV2 layers for building large-scale structures and ResNet blocks for refining small-scale elements in the images. MRI and PET input data are separated out into two elements (base and detail) and then combined using a hybrid technique that combines both mean and maximum feature information. The decoder reconstructs the combined image and a hybrid loss function is used to preserve structural integrity and visual quality of the combined image. In tests using benchmark datasets, FusionNeXt has produced images with clearer, more distinct edges, better contrast, and more details than the original images. Therefore, we believe that FusionNeXt has strong potential to be a tool for better diagnoses and clinical decision making.

Fields of Study & Subjects

Advanced Image Fusion Techniques Medical Imaging Techniques and Applications Multimodal Machine Learning Applications

Publication details

Journal / source
IEEE International Students' Conference on Electrical, Electronics and Computer Science (SCEECS)
Pages
1–6
Publication date
31 January 2026
Language
EN