Michael Sapienza

University of Malta, University of Oxford

Papers

Total Citations

H-Index

About

Michael Sapienza is a computer vision researcher whose work bridges the critical gap between offline batch processing and real-time, online visual intelligence. His primary research areas include human action detection, instance segmentation, and active vision systems for robotics. Sapienza’s most significant contribution is the development of "Incremental Tube Construction" for human action detection, a method that enables systems to recognize actions in streaming video rather than requiring the entire video upfront—a breakthrough essential for applications like human-robot interaction. This work, published in 2017 and 2018, has garnered foundational citations in the field. He also advanced real-time computer vision with "Straight to Shapes++," an extension that made instance segmentation both more accurate and fast enough for deployment in autonomous driving and robotic manipulation. Earlier, his work on the "Real-time visuomotor update of an active binocular head" demonstrated how visual feedback can dynamically control robotic vision systems. Through these contributions, Sapienza has helped push computer vision from static analysis toward the dynamic, real-time perception needed for interactive and autonomous systems.