2023
DOI: 10.1088/2516-1091/acc2fe
|View full text |Cite
|
Sign up to set email alerts
|

Deep multimodal fusion of image and non-image data in disease diagnosis and prognosis: a review

Abstract: The rapid development of diagnostic technologies in healthcare is leading to higher requirements for physicians to handle and integrate the heterogeneous, yet complementary data that are produced during routine practice. For instance, the personalized diagnosis and treatment planning for a single cancer patient relies on various image (e.g., radiological, pathological and camera image) and non-image data (e.g., clinical data andgenomic data). However, such decision-making procedures can be subjective, qualitat… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

1
7
0

Year Published

2023
2023
2024
2024

Publication Types

Select...
6
4

Relationship

0
10

Authors

Journals

citations
Cited by 49 publications
(8 citation statements)
references
References 109 publications
1
7
0
Order By: Relevance
“…By fusing multiple degraded images, DCNN-based methods can effectively de-noise and enhance the image quality, enabling better visualization and interpretation of medical conditions. These applications demonstrate the versatility and effectiveness of DCNN-based medical image fusion techniques in improving image quality, accuracy, and clinical decision-making in various healthcare scenarios [14].…”
Section: Introductionmentioning
confidence: 77%
“…By fusing multiple degraded images, DCNN-based methods can effectively de-noise and enhance the image quality, enabling better visualization and interpretation of medical conditions. These applications demonstrate the versatility and effectiveness of DCNN-based medical image fusion techniques in improving image quality, accuracy, and clinical decision-making in various healthcare scenarios [14].…”
Section: Introductionmentioning
confidence: 77%
“…OpenAI has released plugins for images and will certainly develop multimodal foundational models in the future (e.g., GPT-4 will have capabilities to process images in the near future). But fairness research indicates that the combination with additional information or modality may not necessarily improve performance and is likely to bring about new unfairness and bias (65). For example, in CLIP, a language-vision model, historical race and gender bias are reinforced (66).…”
Section: One Lie Leads To Another: Bias Perpetuationmentioning
confidence: 99%
“…In the current big data era [20] (see review for public neuroimaging datasets [19,100]), how to fully utilize multimodal information is also a significant challenge. Multimodal analysis techniques based on deep learning have made great progress [101], in which techniques such as reinforcement learning [102] and multimodal generalized foundation models [103,104] can be combined with brain parcellations to achieve better performance via the fusion of complementation information across multimodal. Importantly, it is expected to explore the interaction between microscopic phenomena, brain functions and the external environment [99], e.g.…”
Section: Hierarchy Dynamic and Multimodalitymentioning
confidence: 99%