{"id":269,"date":"2023-12-18T02:25:03","date_gmt":"2023-12-18T02:25:03","guid":{"rendered":"https:\/\/mscvprojects.ri.cmu.edu\/f23team9\/?page_id=269"},"modified":"2023-12-18T02:56:11","modified_gmt":"2023-12-18T02:56:11","slug":"fall23","status":"publish","type":"page","link":"https:\/\/mscvprojects.ri.cmu.edu\/f23team9\/fall23\/","title":{"rendered":"Resources"},"content":{"rendered":"\n

Links<\/strong><\/h2>\n\n\n\n

Our project has been wrapped up into a publication. Code and paper could be found in our project page<\/a>. <\/p>\n\n\n\n

\n
Project Page<\/a><\/div>\n\n\n\n
Poster<\/a><\/div>\n<\/div>\n\n\n\n

Summary<\/strong> – Fall 2023<\/strong><\/h2>\n\n\n\n

In this semester:<\/p>\n\n\n\n

    \n
  1. We developed a lightweight plug-in module, the amodal expander, to transform standard, modal trackers into amodal ones through fine-tuning on a few hundred video sequences with data augmentation. <\/li>\n\n\n\n
  2. We introduced PasteNOcclude, a data augmentation technicuqe for crafting occlusion scenarios, that benefits the model in perceiving occluded objects. <\/li>\n\n\n\n
  3. Amodal Expander trained along with PasteNOcclude achieves a 3.3% and 1.6% improvement on the detection and tracking of occluded objects on TAO-Amodal. When evaluated on people, our method produces dramatic improvements of 2x<\/strong> compared to state-of-the-art modal baselines.<\/li>\n<\/ol>\n\n\n\n

    Summary<\/strong><\/h2>\n\n\n\n