{"id":81,"date":"2020-12-16T21:19:23","date_gmt":"2020-12-16T21:19:23","guid":{"rendered":"http:\/\/mscvprojects.ri.cmu.edu\/2020teamm\/?page_id=81"},"modified":"2020-12-17T02:05:23","modified_gmt":"2020-12-17T02:05:23","slug":"fall-semester-2","status":"publish","type":"page","link":"https:\/\/mscvprojects.ri.cmu.edu\/2020teamm\/fall-semester-2\/","title":{"rendered":"Fall ’20 Semester"},"content":{"rendered":"\n

Dataset <\/strong><\/h2>\n\n\n\n

We have one capture that we train on and 4 captures for testing. The first row of the the table shows the factors that are common between the train capture and each of the test capture. The second and third rows show how the appearances of the subject differ between the captures.<\/p>\n\n\n\n

<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Proposed Methods<\/strong><\/h2>\n\n\n\n

Metric Learning<\/h4>\n\n\n\n
\"\"
Diagram for Metric Learning<\/figcaption><\/figure>\n\n\n\n

In this semester, we explored Metric Learning. The idea was to learn to transform input images and predicted texture to a generic feature space. We wanted a comparison in this feature space to minimize the distance between current code (3-view) to 11-view results and at inference, use this transformation to \u2018refine\u2019 the code using Gradient descent. This relies heavily on the assumption that 3-views have enough information to arrive at code 11-view. The diagram can be seen in figure. We want to ensure that loss is quadratic for fast gradient descent.<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Mimicking the 11 View Landscape <\/h4>\n\n\n\n
\"\"<\/figure>\n\n\n\n

In this method, instead of ensuring that loss landscape is quadratic, we mimic the loss landscape of 11 view and the now it takes more time to gradient descent to correct expression and the losses are as follows. <\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

Results<\/strong><\/h2>\n\n\n\n

Metric Learning<\/h4>\n\n\n\n

In this method, instead of ensuring that loss landscape is quadratic, we mimic the loss landscape of 11 view and the now it takes more time to gradient descent to correct expression and the losses are as follows. <\/p>\n\n\n\n