Beyond the Limits of the Human Eye
Petview enables clinicians to visually access information in standard diagnostic images that simply can not be seen with the human eye. Our patented algorithms identify patterns at the pixel level, and then re-visualize them in color so they are easily visible. Now, for the first time, we can see tissue characteristics that provide intricate details that may be used by the clinician to help differentiate disease conditions, including the margins between normal and abnormal tissues.
Cancer has predictable characteristics
It turns out that various types of tissue have tell tale fractal patterns. As PetView’s algorithms transform unstructured image data into structured image data, diffuse grayscale and dense white areas emerge as discernible and quantitatively measurable patterns. The fractal patterns for normal tissue has certain characteristics which are quite different than those for normal or benign tissue. Many of these differences can be seen by the clinician, others can only be detected by using artificial intelligence and machine learning models.
Real-time, actionable analytics
PetView has created machine learning models that take full advantage of the intricate fractal patterns revealed by our patented visual algorithms. While all the other artificial intelligence models we are aware of use the grayscale pixel information in standard diagnostic images, PetView is the only provider who first enhances the source images before building machine learning models. We call this “pre-processing” and it provides several advantages to our models:
- Structures the pixel data – converting the unstructured grayscale pixel data to fractal patterns organizes the data such that it can be better differentiated by the models.
- More usable pattern data – conversion to fractal patterns provides a richer patterns for differentiation.
- Requires fewer training images – because more unique feature data can be accessed for each image, reliable PetView models can be built using a fraction of the total number of images required for typical machine learning models.
Better Clinical Outcomes
The Petview portal is a secured web site accessible using a standard internet connection. The clinician uploads standard images via the web interface, and within about a minute, the following is returned.
- New visualization for each image, showing tissue details and structures not visible in the original.
- Predictive analytics that categorize tissue in each image as normal, benign or malignant.
With this new information available within minutes, with the pet owner still in the clinic, clinicians have visual and predictive analytics to support whatever decision is made.
- Better diagnoses
- Better biopsy decisions
- Better treatment decisions
- Better referral decisions