English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Background and Objective: In recent years, Artificial Intelligence (AI) andin particular Deep Neural Networks (DNN) became a relevant research topic inbiomedical image segmentation due to the availability of more and more datasets along with the establishment of well known competitions. Despite thepopularity of DNN based segmentation on the research side, these techniques arealmost unused in the daily clinical practice even if they could supporteffectively the physician during the diagnostic process. Apart from the issuesrelated to the explainability of the predictions of a neural model, suchsystems are not integrated in the diagnostic workflow, and a standardization oftheir use is needed to achieve this goal. Methods: This paper presents IODeep anew DICOM Information Object Definition (IOD) aimed at storing both the weightsand the architecture of a DNN already trained on a particular image datasetthat is labeled as regards the acquisition modality, the anatomical region, andthe disease under investigation. Results: The IOD architecture is presentedalong with a DNN selection algorithm from the PACS server based on the labelsoutlined above, and a simple PACS viewer purposely designed for demonstratingthe effectiveness of the DICOM integration, while no modifications are requiredon the PACS server side. Also a service based architecture in support of theentire workflow has been implemented. Conclusion: IODeep ensures fullintegration of a trained AI model in a DICOM infrastructure, and it is alsoenables a scenario where a trained model can be either fine-tuned with hospitaldata or trained in a federated learning scheme shared by different hospitals.In this way AI models can be tailored to the real data produced by a Radiologyward thus improving the physician decision making process. Source code isfreely available at https://github.com/CHILab1/IODeep.git

IODeep: an IOD for the introduction of deep learning in the DICOM standard