Leonardo Tanzi

Leonardo Tanzi, Ph.D.

I am currently a Deep Learning Engineer at Methinks AI, a Barcelona start-up focusing on stroke medical assistance. I obtained my Ph.D. in the 3DLab of the Polytechnic University of Turin, Italy, spending a period of research in Paris at UTC Sorbonne Alliance Université under the supervision of Prof. Yves Grandvalet.

I’ve received a Master’s Degree in Computer Engineering, Graphic and Multimedia Curricula, in 2019, after a six-month thesis research experience at KTH University of Stockholm, Department of Technology and Health.

My Ph.D. research focused on human-machine methodologies for smart support during complex interventions, using Deep Learning and Computer Vision for 3D assistance, with a focus on medical application for computer-assisted surgery during pre-operative and intra-operative procedures. I have collaborated with several research groups in the medical field, such as the CTO of Turin’s orthopedic team, the San Luigi Hospital’s urology team, and the Molinette’s maxillofacial team. My main projects are broadly discussed here.

I have lived in four countries (Italy, Sweden, France, and Spain) and speak English, French, Spanish, and Italian.

AI Technical Writer

I love to write and discuss stuff related to AI, in my Medium, I describe and explain different paper that I find exciting for the research community. In addition, I’m collaborating with MarkTechPost, a California-based Artificial Intelligence news platform which focuses on spreading AI Awareness across the globe, reaching more than 1M views per month.

Teaching

I had the opportunity to teach a part of the course 3D Biometric Application, a Master’s Course at Polytechnic University of Turin, a course that aims to provide an understanding of 3D tools and methods for acquiring, modeling, visualizing, and managing data from human anatomy to develop intelligent applications in different application fields. My lessons were focused on an Introduction to Machine Learning and Deep Learning. You can find the slides that I prepared and used here.

Lectures

In these three years of Ph.D., I had the chance to give e guest lecture at the IEEE Fiji for the University of South Pacific, related to the applications of Transformers in vision. I am also part of the community of Datacraft Paris, a learning & coworking club for data scientists, which allows data scientists and data engineers to share best practices and train with their peers. In this contest, I had the chance to give a seminar on Vision Transformer for femur fractures classification.

Generative Adversarial Art

In my free time, I am collaborating with a collective of artists based in Turin, in the context of applying AI and in particular Generative Adversarial Network to image and text-to-image synthesis. As a first projects, we produced a musical video based on the lyrics, based on text-to-image Diffusion Models, here. Some preliminary results of generative art can be found here.

Scientific Publications

  1. Vision Transformers for femur fracture classification. Tanzi L, Audisio A, Cirrincione G, Aprato A, Vezzetti E. Injury (2022). DOI
  2. Real-time deep learning semantic segmentation during intra-operative surgery for 3D augmented reality assistance. Tanzi L, Piazzolla P, Vezzetti E, Porpiglia. International Journal of Computer Assisted Radiology and Surgery (2021). DOI
  3. Hierarchical Fracture Classification of Proximal Femur X-Ray Images Using a Multistage Deep Learning Approach. Tanzi L, Vezzetti E, Moreno R, Aprato A, Audisio A, Massè A. European Journal of Radiology 133 (2020). DOI
  4. Intraoperative Surgery Room Management: a Deep Learning Perspectives. Tanzi L, Piazzolla P, Vezzetti E. The International Journal of Medical Robotics and Computer Assisted Surgery (2020). DOI
  5. A deep learning framework for real-time 3D model registration in robot-assisted laparoscopic surgery. Padovan E, Marullo G, Tanzi L, Piazzolla P, Moos S, Porpiglia F, Vezzetti E. The International Journal of Medical Robotics and Computer Assisted Surgery (2022). DOI
  6. X-Ray Bone Fracture Classification Using Deep Learning: a Baseline For Designing a Reliable Approach. Tanzi L, Vezzetti E, Moreno R, Moos S. Applied Sciences (2020). DOI
  7. Exploiting deep learning and augmented reality in fused deposition modeling: a focus on registration. Tanzi L, Piazzolla P, Moos S, Vezzetti E. International Journal on Interactive Design and Manufacturing (2022). DOI
  8. 6D object position estimation from 2D images: a literature review. Marullo G, Tanzi L, Piazzolla P, Vezzetti E. Multimedia Tools and Applications (2022). DOI
  9. Dynamic evaluation of THA components by Prosthesis Impingement Software (PIS). Giachino M, Aprato A, Revetria TA, Vezzetti E, Massè A, Ulrich L, Tanzi L. Acta Biomededica (2021). DOI
  10. Advanced deep learning comparisons for non-invasive tunnel lining assessment from ground penetrating radar profiles. Rosso M, Marasco G, Tanzi L, Aiello S, Aloisio A, Cucuzza R, Chiaia B, Cirrincione G and Marano G. Eccomas 2022. DOI