Genki highlights regions of a chest X-ray that have specific abnormalities associated with tuberculosis (TB) and helps radiologists to prioritize and reduce their workload. Genki can be an adaptive system that, if deployed online, evolves over time based on every new X-ray read in the field. Along with displaying the heat map, outline and bounding boxes, it also automatically generates a radiology report compliant with Radiological Society of North America standards. A human-verified signed report from a board certified radiologist can also be made available if required.

Demo site available on request. Please email info@deeptek.ai.

Certification

CE (pending), FDA (pending)

Development Stage

On the market

Target Setting

Primary health centres, general hospitals, teleradiology
centres, large-scale government-run TB screening programmes, etc.

Deployment

Online & offline

Intended Age Group

14+ years

Current Market

India, Asia-Pacific region and Africa.

Future market (after CE/FDA certification): USA, Europe and Japan.

Input

High quality posterior-anterior or anterior-posterior chest X-ray image in DICOM/PNG/JPG formats

Output

Structured report including:

»» Heat map showing regions suspected of TB

»» Probability score for TB

»» Probability scores for 20+ pulmonary abnormalities other than TB

Hardware

i5 processor or equivalent, 8 GB RAM, 1 TB hard-drive (for local installation)

Validation

The product has been internally and externally validated by several researchers, customers, and healthcare foundations.

Hardware

i5 processor or equivalent, 8 GB RAM, 1 TB hard-drive (for local installation)

Server

Both cloud deployment and local on-premise deployment are supported.

Integration

Integration with existing picture archiving and communication system (PACS) supported.

Software

Cloud deployment: Browser (Chrome, Firefox)
Local deployment: Microsoft Windows / Linux / Docker (all other dependencies will be handled by Docker)

Processing Time

Less than 30 seconds

Data Sharing & Privacy

»» Server location: both cloud deployment and local on-premise deployment are supported.

»» Patient scans are uploaded to the cloud after anonymization and not used for training without patient consent.

»» Data are anonymized before processing.

Price

Volume-based pricing models are available. Please contact info@deeptek.ai.

Software Updates

»» Minor updates monthly (for online only), major updates yearly (for online and offline)

»» Software upgrades are included in the license price

Product Development Method

Supervised and unsupervised deep learning model architectures

Training

The artificial intelligence models were trained on 1.5 million chest X-rays from over 10 countries.

Reference Standard

High-quality radiologist-adjudicated ground truth

Publications

Please visit https://www.deeptek.ai/publications