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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Weakly supervised disease classification of CT imaging suffers from poorlocalization owing to case-level annotations, where even a positive scan canhold hundreds to thousands of negative slices along multiple planes.Furthermore, although deep learning segmentation and classification modelsextract distinctly unique combinations of anatomical features from the sametarget class(es), they are typically seen as two independent processes in acomputer-aided diagnosis (CAD) pipeline, with little to no feature reuse. Inthis research, we propose a medical classifier that leverages the semanticstructural concepts learned via multi-resolution segmentation feature maps, toguide weakly supervised 3D classification of chest CT volumes. Additionally, acomparative analysis is drawn across two different types of feature aggregationto explore the vast possibilities surrounding feature fusion. Using a datasetof 1593 scans labeled on a case-level basis via rule-based model, we train adual-stage convolutional neural network (CNN) to perform organ segmentation andbinary classification of four representative diseases (emphysema,pneumonia/atelectasis, mass and nodules) in lungs. The baseline model, withseparate stages for segmentation and classification, results in AUC of 0.791.Using identical hyperparameters, the connected architecture using static anddynamic feature aggregation improves performance to AUC of 0.832 and 0.851,respectively. This study advances the field in two key ways. First, case-levelreport data is used to weakly supervise a 3D CT classifier of multiple,simultaneous diseases for an organ. Second, segmentation and classificationmodels are connected with two different feature aggregation strategies toenhance the classification performance.

Weakly Supervised 3D Classification of Chest CT using Aggregated Multi-Resolution Deep Segmentation Features