<div dir="ltr">Hi All,<div> A reminder that our kick-off reading group session is on tomorrow, Friday at 12pm in the 2nd floor meeting room in CNSI (Room 2003).</div><div><br></div><div>Abhishek from Mirage Lab will lead this first discussion. Below is the CVPR paper that he will present (and related blog posts) from the reading-group slack channel. </div><div><br></div><div>Hope to see you there!</div><div><br></div><div>-John</div><div><br></div><div>------</div><div><br></div><div>Hi everyone, I will be presenting the following paper on Friday:</div><div><br></div><div><a href="https://people.eecs.berkeley.edu/~tinghuiz/projects/SfMLearner/">https://people.eecs.berkeley.edu/~tinghuiz/projects/SfMLearner/</a> (CVPR 2017 Oral)</div><div><br></div><div>This paper falls under a broad category of unsupervised 3D deep learning approaches for single image depth estimation.</div><div>The following blog posts give a good motivation and high level overview:</div><div><br></div><div><a href="http://bair.berkeley.edu/blog/2017/07/11/confluence-of-geometry-and-learning/">http://bair.berkeley.edu/blog/2017/07/11/confluence-of-geometry-and-learning/</a></div><div><br></div><div><a href="https://alexgkendall.com/computer_vision/Reprojection_losses_geometry_computer_vision/">https://alexgkendall.com/computer_vision/Reprojection_losses_geometry_computer_vision/</a></div><div><br></div><div>I will be mainly focusing on the above paper. If time permits I will also breifly talk about the following related papers:</div><div><br></div><div><a href="https://arxiv.org/abs/1609.03677">https://arxiv.org/abs/1609.03677</a> (CVPR 2017 Oral)</div><div><br></div><div><a href="https://arxiv.org/abs/1702.02706">https://arxiv.org/abs/1702.02706</a> (CVPR 2017 Spotlight Oral)</div><div><br></div><div><a href="https://shubhtuls.github.io/drc/">https://shubhtuls.github.io/drc/</a> (CVPR 2017 Oral) </div><div><br></div><div><br></div><div><br></div><div><br clear="all"><div><br></div>-- <br><div class="gmail_signature"><div dir="ltr"><div><div dir="ltr">John O'Donovan<br><div>Department of Computer Science<br>University of California, Santa Barbara, CA 93106-5110<br><br>email: <a href="mailto:jod@cs.ucsb.edu" target="_blank">jod@cs.ucsb.edu</a><br>phone: (805)451-9342<div>web: <a href="http://cs.ucsb.edu/~jod" target="_blank">http://cs.ucsb.edu/~jod</a></div></div></div></div></div></div>
</div></div>