<html><head><meta http-equiv="content-type" content="text/html; charset=us-ascii"></head><body style="overflow-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;"><div>This is a reminder of Joshua Ahn's MS Presentation</div><div><br></div><div>===============================================</div><div><br></div><div>Candidate: <span style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0);">Joshua Ahn</span></div><div><br></div><div>Date: Friday, April 12, 2024</div><div><br></div><div>Time: 4 pm CT</div><div><br></div><div>Zoom Link: <a href="https://uchicago.zoom.us/j/9040721508?pwd=amovMGJDSW83R2JkVC9FMVMydUttQT09">https://uchicago.zoom.us/j/9040721508?pwd=amovMGJDSW83R2JkVC9FMVMydUttQT09</a></div><div><br></div><div>Location: JCL 356</div><div><br></div><div>Title: On Alpha Invariance and the Inverse Scaling between Distance and Volume Density in Neural Radiance Fields</div><div><br></div><div>Abstract: Scale-ambiguity in 3D scene dimensions leads to magnitude-ambiguity of volumetric densities in neural radiance fields, i.e., the densities halve when the scene size doubles, and vice versa. We call this property alpha invariance. For NeRFs to better maintain alpha invariance, we recommend 1) parameterizing both distance and volumetric density in log space, and 2) using a discretization-agnostic initialization strategy to guarantee high transmittance and consistently high rendering quality. We revisit several popular radiance field models from the literature and find that these architectures use various heuristics to address issues arising from scene scaling. We test their behaviors and demonstrate that our recipe is more robust to changes in scene size.</div><div>I will send you a draft of the paper closer to next Friday. Thank you!</div><div><br></div><div>Advisors: Greg Shakhnarovich</div><div><br></div><div>Committee Members: Greg Shakhnarovich, Rana Hanocka, Michael Maire</div><div><br></div><div>
<div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;"><div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;"><div dir="auto" style="caret-color: rgb(0, 0, 0); color: rgb(0, 0, 0); letter-spacing: normal; text-align: start; text-indent: 0px; text-transform: none; white-space: normal; word-spacing: 0px; -webkit-text-stroke-width: 0px; text-decoration: none; word-wrap: break-word; -webkit-nbsp-mode: space; line-break: after-white-space;"><div><div><font color="#610003" face="Avenir Next"><b>Jessica Garza</b></font><br><font face="Avenir-Book">Assistant Director of Undergraduate Studies</font><br><font face="Avenir-Book">Department of Computer Science</font></div><div><font face="Avenir-Book">The University of Chicago</font></div><div><font face="Avenir-Book">John Crerar Library 374</font></div><div><font face="Avenir-Book">Office: (773) 702-2336</font></div></div></div></div></div>
</div>
<br><div><br></div><div><br></div><div></div></body></html>