������

Finding the dental nerve root canal and determining its size is usually time-consuming. Now, it can be automatically identified through machine and deep learning methods, making the work of healthcare professionals faster and easier, in addition to making the treatment safer.

The project grew out of cooperation with FCAI and from the very start involved clinicians from Tampere University Hospital. “They also greatly influenced the choice of research topics. It was important for us to that the topics were selected so that they will have real clinical impact. Segmentation of the mandibular nerve canal and marking of cephalometric points in 3D images were selected as research topics”, says Research Manager Ari Hietanen from Planmeca.

Innovating in the healthcare space requires knowledge and skills from many domains, which is why the three-way collaboration of FCAI, Planmeca and Tampere University Hospital makes for a winning combination. The necessary data and clinical expertise come from the hospital clinicians, methodology research and expertise from the university, and technical application and business expertise from the company.

Planmeca has been one of FCAI's original partners since May 2018. The funding received through this cooperation has especially supported Jaakko Sahlsten, a dissertation researcher with Professor Kimmo Kaski, and Joel Jaskari, a dissertation researcher with Professor Simo Särkkä. They have developed deep learning methods for the project and actively published their results. “Publications constitute an important part of the collaborative project. They advance the theses of PhD students and provide critical information for the medical experts”, says Sahlsten.

For Planmeca, the collaboration means new and significant business opportunities. “Our main products are digital X-ray machines and dental units. It has always been important for us to be at the forefront of development, and we have invested heavily in research and development. Machine learning solutions are a natural extension of this, as they will be an important part of future technology solutions”, says Hietanen.

Reference: Jaskari, J., Sahlsten, J., Järnstedt, J. et al. . Sci Rep 10, 5842 (2020).