Around 10% of population in developed countries suffer a stone episode at least once in his/her life. This disease has a high recurrence rate, around 40% in 5 years. Once the stone episode has passed, it is widely agreed that an adequate study of the causes of stone formation is required in order to decrease the recurrence probability.

myStone is a new device and software for capturing images of expelled kidney stones and automatically classify them in a cheap, fast and on-site way. The classifier is trained with a dataset of thousands of images of stone 900 fragments, in the visible and near infrared spectrum.

The user shows the system the inner and external views of two fragments, obtained from extracorporeal shock wave lithotripsy or naturally expelled.


The software outputs a probability for each class and also generates a report with recommendations for the urologist or diet advises for the patient. This overcomes the lack of expensive equipment and expert technicians, time delay and, to some extent, poor classification specificity.

Generated report

To know more go to the project site




Felipe Lumbreras, Joan Serrat, Fran Blanco, Natali, Gemma Roig, Victor Garcia, Montserrat López, Manuel Valiente, Albert Pell