Behold.ai is deep-learning medical software that makes it easy for healthcare practitioners to identify diseases from radiology and camera image data. Similar to the way Facebook recommends tags for friends in your pictures, behold.ai recommends tag for abnormalities in medical images. Doctors can then follow the recommendation by adding or removing tags. It was developed to enable the community of healthcare workers to leverage the existing medical data in conjunction with artificial intelligence to perform large scale visual recognition. It utilizes deep-learning technology, which combines convolutional neural network architectures with massive computing power to allow software programs to identify visual differences between healthy and an unhealthy body parts. While tech giants like Google, Facebook and Amazon invest heavily in this technology for consumer applications and ecommerce, behold.ai brings this power to the healthcare sector. It delivers an inexhaustible way to automati...
Behold.ai is deep-learning medical software that makes it easy for healthcare practitioners to identify diseases from radiology and camera image data. Similar to the way Facebook recommends tags for friends in your pictures, behold.ai recommends tag for abnormalities in medical images. Doctors can then follow the recommendation by adding or removing tags. It was developed to enable the community of healthcare workers to leverage the existing medical data in conjunction with artificial intelligence to perform large scale visual recognition. It utilizes deep-learning technology, which combines convolutional neural network architectures with massive computing power to allow software programs to identify visual differences between healthy and an unhealthy body parts. While tech giants like Google, Facebook and Amazon invest heavily in this technology for consumer applications and ecommerce, behold.ai brings this power to the healthcare sector. It delivers an inexhaustible way to automatically diagnose images with nodules, lesions and other abnormalities.

We ingest images from a variety of medical imaging techniques, such as: dermoscopy, mammograms, CT scans, retinal scans and magnetic resonance imaging (MRIs), and then perform a preliminary predictive diagnosis. A physician could then use the prediction to make a quicker decision on how they think the patient should undergo treatment. The training dataset grows larger with usage allowing behold.ai to learn overtime thereby increasing the accuracy and consistency of predictions. The software’s ability to perform pattern recognition will greatly reduce the time that a radiologist would have to spend looking at an image, saving both hospitals money, and allowing patients to receive treatment quicker.

The technology works by first taking, as input, an image from an imaging machine such as a CT scan. Secondly, the image is processed by the image recognition algorithms that have been trained to recognize abnormalities in specific image datasets. Thirdly, the software outputs its predictions along with the percentage likelihood of diseases. Medical practitioners would then decide if their predictions match the machine and determine a possible course of treatment.
More information

Investors

F6S  Alpha
Admin
gener8tor Madison 2016
Admin
gener8tor Madison 2016 A Concierge Startup Accelerator