AI Science Tracker
EN

University of Azad Jammu and Kashmir

1 discovery from 2021

Discoveries

Small clinical CT-image check reveals COVIDC classification errors

Source check: Source checked. Not independent scientific confirmation.

Finding: In a radiologist-supervised hospital check using scans excluded from training, COVIDC classified9 of10 supplied COVID-labeled images and8 of10 non-COVID images correctly. A separate20-image severity check classified19 correctly. This small selected image-level reference evaluation does not establish patient-level diagnosis, prognosis or clinical utility.

Technology: Pretrained CNN feature extraction; support vector classifier

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
ImageNet-pretrained CNN feature maps feed learned support-vector classifiers for supplied COVID/non-COVID and severe/mild image labels.
Test and result
After development on public CT images, the source reports20 raw hospital images for COVID classification and another20 for severity, not included in training and evaluated under radiologist supervision. The observed confusion counts are retained rather than the near-perfect development benchmark.
Limitation / still unknown
The source does not establish unique patient counts, PCR/reference-label ascertainment, blinded case selection or independence between the two small hospital sets. Development cross-validation shuffled images rather than demonstrating patient-grouped folds. Selecting model variants on development comparisons limits generalization; supplied severity labels are not future deterioration outcomes. No quarantine, intensive-care recommendation or patient benefit is adopted.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC7901302/ · https://doi.org/10.1016/j.imu.2021.100540

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Original clinical dataset, pretrained feature/classifier Methods, cross-validation and separate AIMS evaluation Results read.

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 1/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open the interactive profile for the timeline, or browse all organizations and model families.