EN
FR AR

AI Portfolio

Hckonnect

Chest x-ray

Abnormal findings detecting in chest X-ray images

Radiology
Hckonnect
Hckonnect

Workflow Improvement

Currently, viewing stations do not allow for intelligent sorting and cases prioritizing – a radiologist can only sort based on name, date of image acquisition, time of image acquisition, etc.

The system analyzes the abnormality of five major findings to support doctors with their diagnosis of major lung diseases such as lung cancer, tuberculosis and pneumonia based on a combination of findings and allows to:

  •  Prioritize cases depending on their severity.   •  Assist diagnosis by providing a contouring of the abnormality

Hckonnect

Easy Integration with Existing Infrastructures

The software maximizes user convenience by delivering information through an intuitive user interface (UI) and providing optimized forms of services such as APIs for user reading system settings.

It offers a cloud-based service and is integrated with PACS.

  •  Provides list sorting and search function (Patient ID, Date, Upload Time).   •  Analyzes multiple images simultaneously.

Hckonnect

Proven Performance via External Validation

VUNO Med- Chest X-ray has been trained on tens of thousands of chest PA X-ray images collected from Korea’s leading hospitals to ensure:

  •  Reduced reading time by more than 50%1(Fig.1)
  •  Improved reading accuracy of radiology residents and radiologists by average of 8% 1*(Fig.2)
  •  0.985 AUROC per-image / 0.943 JAFROC FOM per-lesion / FP 0.12

Publications

Added Value of Deep Learning–based Detection System for Multiple Major Findings on Chest Radiographs
https://id.elsevier.com/as/authorization.oauth2/html

Deep Learning–based Differentiation of Invasive Adenocarcinomas from Preinvasive or Minimally Invasive Lesions among Pulmonary Subsolid Nodules
https://tlcr.amegroups.org/article/view/49486/html

Related products

View all products