Carlo Tomasi

Tomasi's research is at the intersection of computer vision, machine learning, and applied mathematics. Tomasi's current projects include image motion analysis (funded by NSF), satellite image interpretation (funded by IARPA), computer-assisted diagnosis, and object recognition (funded by Amazon). He is an ACM Fellow and has won the IEEE Computer Society Helmholtz Prize twice.

Current Appointments & Affiliations

Iris Einheuser Distinguished Professor · 2016 - Present Computer Science, Trinity College of Arts & Sciences

Professor of Computer Science · 2004 - Present Computer Science, Trinity College of Arts & Sciences

Faculty Network Member of the Duke Institute for Brain Sciences · 2011 - Present Duke Institute for Brain Sciences, University Institutes and Centers

Recent Publications

Cross-Attention Transformer for Video Interpolation

Conference Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) · January 1, 2023 We propose TAIN (Transformers and Attention for video INterpolation), a residual neural network for video interpolation, which aims to interpolate an intermediate frame given two consecutive image frames around it. We first present a novel vision transform ... Full text Cite

SemARFlow: Injecting Semantics into Unsupervised Optical Flow Estimation for Autonomous Driving

Conference Proceedings of the IEEE International Conference on Computer Vision · January 1, 2023 Unsupervised optical flow estimation is especially hard near occlusions and motion boundaries and in low-texture regions. We show that additional information such as semantics and domain knowledge can help better constrain this problem. We introduce SemARF ... Full text Cite

Optical Flow Training Under Limited Label Budget via Active Learning

Conference Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) · January 1, 2022 Supervised training of optical flow predictors generally yields better accuracy than unsupervised training. However, the improved performance comes at an often high annotation cost. Semi-supervised training trades off accuracy against annotation cost. We u ... Full text Cite

View All Publications

External Links

Website