Heartwise

Most AI models for electrocardiogram interpretation require access to the raw digital waveform, which is only available inside hospital systems with full digital infrastructure. In practice, the vast majority of ECGs around the world exist as printed reports, scanned PDFs, or photos taken on a phone, which makes current AI tools unusable in those settings.

With TICAN’s support, Heartwise is training a vision foundation model that interprets ECGs directly from images, without needing the underlying signal. The same model can then be used for automated diagnosis, risk prediction, and downstream clinical decision support from a simple smartphone photo or scan, extending advanced ECG AI to clinics and regions that do not have access to integrated digital ECG systems.