Harrison.ai (annalise.ai) has introduced Harrison.rad.1, a new vision language model for radiology, trained on diverse clinical data, including millions of radiology images and reports. This single model can perform a diverse set of tasks such as X-ray interpretation, findings localization, and report generation with high accuracy.
Currently, Harrison.rad.1 is available to select industry partners and healthcare professionals, to gain a better understanding of the capabilities and limitations of these models, and to determine if these models truly improve quality and save time in clinical workflows. As we have not seen models like this before in radiology practice it will be interesting to learn what impact they have and how patient safety is being warranted.
Read full press release