ASSISTIVE DIAGNOSIS AI TOOLS

Axial AI (uAI Discover PNA), an Artificial Intelligence (AI)-based COVID-19 AI-assisted CT Scan diagnosis platform is a joint development project led by Shanghai Research Centre for Brain Science and Brain-Inspired Intelligence, United Imaging – United Imaging Intelligence and Skymind Neurobionix Research.

flipbox-image

Learn who are the experts behind Axial AI

About us

Assistive Diagnosis AI Tools for COVID-19

Axial AI (uAI Discover PNA), an Artificial Intelligence (AI)-based COVID-19 AI-assisted CT Scan diagnosis platform is a joint development project led by Shanghai Research Centre for Brain Science and Brain-Inspired Intelligence, United Imaging – United Imaging Intelligence and Skymind Neurobionix Research.

Axial AI (uAI Discover PNA) was developed to assist the medical teams to diagnose patients with COVID-19 symptoms faster. It can automate the analysis of CT Scan images of possible COVID-19 patients within 10 seconds, with an accuracy of more than 90%.

The system was extensively used in the Huoshenshan Hospital, an emergency specialty field hospital constructed from 23 January 2020 to 2 February 2020 in response to the 2019–20 coronavirus pandemic, and hospitals throughout Wuhan, Hubei and other provinces in China to assist with the diagnosis and treatment of COVID-19.

Its capability to provide precise quantitative data has assisted medical practitioners to identify the location of infection faster hence providing them opportunity to give the accurate treatment to the patients. This has proven to help the doctors and shown tremendous success recovery rate.

Axial AI Main Features

AI-TRIAGE

AI-CLASSIFICATION

AI-ANALYSIS

AI FOLLOW-UP

AI-REFERENCE

The user interface

ai-triage

AI-Classification of Viral Pneumonia

ai-CLASSIFICATION

Fast & Robust Segmentation and Quantification

ai-ANALYSIS

Based on 6000+ Training Dataset: 2000+ positive cases and 3000+ negative cases

Dice Score 90% +
Sensitivity 95%+
Specificity 95%+
Accuracy 95%+

Volumetric Error < 5%

More in-depth of AI-Analysis

ai-analysis-1
ai-analysis-2
ai-analysis-3

Sensitivity

AI-ANALYSIS is sensitive to Multiple Radiographic Patterns

Quantification

Quantification of Infections at Lung Lobe / Segment Level

Analysis

Tissue Type Analysis of the Infected Region using AI-ANALYSIS

AI-based system for patient follow up

ai-FOLLOW-UP

Same Patient Multiple CT Scans

Follow-up scans every 3-5 days are often recommended for COVID-19 patients.

Same Patient Multiple CT Scans

  • Automatic Image Registration & Synchronized Scrolling of Images
  • Comparative Analysis of Different Time-points
  • Structured Follow-up Reports

Providing Similar Patients to Help Diagnosis

ai-REFERENCE

Diagnosis Reports

Both sides of the thorax were symmetrical with no obvious bone abnormalities. The trachea and main bronchi were unobstructed. The posterior and posterior segments of the upper lobe of the right lung, the middle and lower lobe of the right lung, the upper lobe of the left lung, and the dorsal and lateral basal segments of the lower lobe of the left lung were seen in sheet-like ground glass density shadows, with thickening of the intralobular and interlobular septum, partly "pavement stones" changed, part of the consolidation, inflated bronchial signs. Slightly enlarged lymph nodes were seen in the mediastinum. A small amount of pericardial effusion. There was no thickening of the pleura on both sides, and a small amount of liquid density was seen in the thoracic cavity on both sides.

  • Providing a database of anonymized typical COVID-19 patients, including diagnosis reports written by senior doctors.
  • Similar patients are retrieved from the database to help diagnosis of the current patient.

What Else Axial AI Can Do

extended capabilities

uAI Discover - Lung

The system can accurately detect lung nodules larger than 3mm. When the number of false positives per scan is 1, the detection sensitivity of early lung cancer reaches up to 95%.

uAI Discover – Rib

The system automatically detects the rib fractures from CT chest scans within seconds and provides innovative animated visualization of local rib fractures for radiologists to further examine.

uAI EasyTriage – ICH

The system is precisely and efficiently segment cerebral hemorrhage regions with deep learning algorithms, localize hemorrhage anatomical regions, classify hemorrhage levels, and generate quantified and structured diagnostic reports.

uAI Discover & LinQ – Brain

The system generates brain parcellation and provides quantitative metrics of 112 sub-structures of the brain in seconds with one click. It automatically identifies and labels high-risk brain atrophy regions and supports quantitative follow-up studies.

uAI Discover – BoneAge

The system automatically assesses a child’s skeletal maturity and predicts the growth potential from child’s hand radiographs. It localizes and analyzes ossification centers of multiple carpal, metacarpal and phalangeal bones, estimates the bone age and generates structured child’s growth report within seconds.

uAI Discover – TPS

The intelligent radiotherapy planning system is based on our self-developed deep learning engine for 3D image segmentation. It improves the segmentation robustness and accuracy compared to atlas-based approaches and can fast contour a target or an organ at risk within CT/MR in seconds.

See Axial AI in Action

Axial AI has been used in over 60 hospitals in China, including Huo Shen Shan Hospital and Lei Shen Shan Hospital in Wuhan, a specialised Covid-19 hospitals that’s built in 10 days. 

Axial AI is jointly developed by:

© 2020 Skymind Holdings Berhad . All rights reserved.