Preliminary diagnosis of kidney stone characteristics performs a key role in taking efficient therapeutic approaches. Current diagnosis method relies on radiologist interpretation of CT scans. Clinicians spend a great deal of time in image interpretation, which places a great challenge on the efficacy of image diagnosis. Our solution can help in interpreting data retrieved from CT scanning images. We therefore see a significant opportunity to enhance patient care, reduce wait times for radiologist interpretation, and improve both patient and doctor outcomes.
Using ground truth based machine learning to interpret CT scans, we develop an intelligent software application capable of determining kidney stone presence, size, aspect ratio and other geometrical features, as well as the probable composition. Our solution offers the advantage of physical and chemical interpretation of the kidney stones using explainable computer vision algorithms. The solution includes:
This technology will advance the presence of AI in the medical field by use of image-guided therapy, both in machine vision and in spectroscopic correlative analysis of kidney stones.