GAIA PAVONI

About me

My research activity focuses on the design of IT tools for underwater environmental monitoring. Between the different topics: automatic recognition and segmentation from the large-area images, human-in-the-loop applications for optical data visualization, labeling and analysis, and image-based 3D reconstructions of underwater habitats.

I’m one of the main developers of TagLab, an AI-powered interactive annotation tool that speeds up image data human labeling and analysis. TagLab supports the detection of change from multi-temporal surveys, the analysis of multi-channel data, and allows the experts to create custom classifiers through a feature called Train-Your-Network.

I'm holding a Postdoctoral Fellowship at the Visual Computing Lab since November 2014. From 2014 to 2016, my activities have been focused on digital technologies for the Cultural Heritage field. From 2012 to 2014, I worked for TiQU-process, a start-up company of 3D scanning and rapid prototyping.

I have a Ph.D. in Information Engineering, a master's degree in Applied Mathematics (Signal and Image Processing), and a Bachelor's degree in Mathematics. 

TagLab: TagLab's main interface: the export of labeled dataset for the training of semantic segmentation models.

Multi-temporal tracking of areas' changes.

Ongoing projects and collaborations

  • My researches about the automatic underwater recognition are supported by the NVIDIA AI Technology Center (NVAITC).
  • I'm a team member of the project "Monitoring program of benthos in the Ross Sea: new non-destructive and machine-learning-based approaches for the analysis of benthos", (PNRA 2018).
  • I collaborate with the Marine Science Institute (University of California Santa Barbara) to the NSF  Moorea III long-term ecological research project.
  • TagLab has been developed with the ecological technical involvement of the Stuart Sandin Lab of the Scripps Institution of Oceanology (University of California, San Diego).

Publications

A complete list of pubblication can be found HERE

Awards

  • 'Another Brick in the Wall: Improving the Assisted Semantic Segmentation of Masonry Walls' wins the best paper at EG GCH 2020 conference

TagLab has been used successfully used in the stratigraphic analysis of masonry.

  • “SEMANTIC SEGMENTATION OF BENTHIC COMMUNITIES FROM ORTHO-MOSAIC MAPS” was awarded as one of the BEST PAPER AWARD at the Second International workshop of Underwater Photogrammetry and Remote Sensing
  • Alchemy in 3D wins the "2015 Digital humanities award" in the "Best Use of DH for Public Engagement" category.
  • Best Paper Award for the paper: "Alchemy in 3D: A Digitization for Journey Through Matter", IEEE Digital Heritage 2015 conference.

How to reach me​​​​​​​

gaia.pavoni@isti.cnt.it

ISTI-CNR, Area della Ricerca
Via G. Moruzzi 1, 56124, Pisa (PI) 
Phone: +39 050 315 2925