Register for free to
view the full innovation profile.
it's an end-to-end solution, starting from training the AI (labeler platform), to analysis (prediction) via web, DICOM or API integration, to concordance / discordance feedback
About
Our Ai model leverages off machine learning and expertise of radiologists .The chest AI :It detects and highlights 20 different pathologies including covid pneumonia on chest x-rays. RADIFY detects the pathologies, highlights the features with bounding boxes and heat maps. It can label 2000 x-rays in 1 minute thus 2000 times faster than a human being. The life threatening cases are automatically prioritized and the findings are pre- populated on a reporting template enabling treatment to be provided instantly. Our Mammogram AI assists in double reading of mammograms and triages the cases according to a score BIRADS which is the breast imaging reporting and database system score. The solution includes a carefully selected, matched and trained AI models that complement each other, allowing for training models independent of other models, inclusion of additional labels without having to retrain any other models, etc. We also use learning transfer in order to expedite launch of new models based on learnings of other models, i.e. detection of mass in chest x-rays is used to train mammography models to leverage of data set .
Key Benefits
The AI model can label 2000 x-rays in 1 minute thus 2000 times faster than a human being. The life threatening cases are automatically prioritized and the findings are pre-populated on a reporting template for doctors ease thus enabling treatment to be provided instantly. It enables doctors to diagnose more images, more consistently and accurately in less time and at a lower cost whilst prioritizing care for patients who need it most. It can reduce chest x-ray turnaround time by 50% and free up the doctors expertise services to concentrate on other complex diseases. The mammogram model is effective in screening for breast cancer which is the number 1 cancer in women globally. Radiology expertise is critical in building a robust platform customized to the end user needs, which is our key differentiator. We are the only company with a complete end to end solution where doctors can interact, augment the AI findings for concordance/discordance. Our algorithm is built based on object (pathology) detection unlike our competitors who use classifier models and natural language processing thus requiring large amounts of data and reports. We, however, leverage the expertise of radiologists who annotate the pathology on our proprietary labeler platform. We require less data for training and are more accurate, validated by radiologists. We are the only company to provide a pediatric imaging AI solution and have gamified our labeler, used by students/universities in preparation for radiology exams, (CPD) continuous performance evaluation.
Applications
Proposition: AI platform is for doctors/ radiologists who want to make more accurate diagnoses faster and cheaper - Segments: Hospitals , casualty ,Occupational health, mobile health, and telemedicine who typically need to manage high volumes of chest x-rays and mammograms for screening of respiratory conditions, breast cancer and covid pneumonia