BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20260308T030000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3 TZNAME:EDT END:DAYLIGHT BEGIN:STANDARD DTSTART:20251102T010000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11 TZNAME:EST END:STANDARD END:VTIMEZONE BEGIN:VEVENT DTSTAMP:20251116T203404Z UID:BB2332F3-3FF9-4140-B32B-A0ED3C26C5F5 DTSTART;TZID=America/New_York:20251113T190000 DTEND;TZID=America/New_York:20251113T213000 DESCRIPTION:[]\n\nHave you ever tried Google’s Live View feature\, where you just point your phone at a building and it instantly figures out exact ly where you are? Or maybe you’ve wondered how action movies pull off th ose cool 3D scenes of the places they’re about to rob? Ever stopped to t hink about how virtual tours and street views come to life?\n\nIn this tal k\, we will explore the various algorithms that are used to create vivid a nd comprehensive 3D scenes from just a handful of images collected from th e internet. The talk will be divided in 3 key sections:\n\nArtifact Mitiga tion\nImages collected from the internet are not always perfect. Some may have memes or text overlaid on them\, while others may have been compresse d multiple times for efficient storage\, resulting in pixelation. Some ima ges might be blurry\, have too much sun exposure\, or be taken at night. T here can also be moving objects like people and cars\, which aren’t need ed and could obstruct the 3D reconstruction of a scene.\n\nIn this section \, we’ll learn about the deep learning algorithms used to remove these k inds of artifacts and transient objects from images.\n\nImage Registration and Geo-localization\nOnce the images have been pre-processed\, the next step is to determine their relative pose with respect to each other. Imagi ne trying to figure out whether a photo was taken from the left side of th e Eiffel Tower or the right\, or whether the person with the camera was 10 0 meters away or just 50 meters. Sometimes\, the images might even come fr om a drone! So how do we place all these different frames into a common re ference frame? The more viewpoints we have of an area\, the more complete our 3D models will be.\n\nIn this section\, we’ll learn how Structure-fr om-Motion (SfM) is used to assign poses to these images. We’ll explore t echniques for using background details to determine pose\, especially when the object of interest looks the same from every angle. And finally\, we ’ll briefly discuss how these images can be geo-localized\; meaning thei r latitude and longitude can be estimated\, even when no GPS information i s available.\n\n3D Reconstruction\nNow that our images are cleaned up and we know their poses\, we can dive into the techniques used to transform th ese images to build a 3D scene. We’ll discuss\, at a high level\, some t raditional 3D reconstruction methods along with more recent AI-based appro aches such as neural rendering and gaussian splatting.\n\nThe applications of these computer vision and deep learning techniques we’ve talked abou t are widespread\, and I hope you’ll start noticing them being used all around you\, whether it’s in mobile phones\, self-driving technology\, b iometric scanners\, delivery robots\, security cameras\, and many other pl aces.\n\nSpeaker(s): Kshitij Minhas\n\nAgenda: \n7:00 PM Networking and An nouncements\n\n7:10 PM Presentation\n\nVirtual: https://events.vtools.ieee .org/m/506731 LOCATION:Virtual: https://events.vtools.ieee.org/m/506731 ORGANIZER:chair@licn.org SEQUENCE:75 SUMMARY:Creating a 3D World from 2D Images URL;VALUE=URI:https://events.vtools.ieee.org/m/506731 X-ALT-DESC:Description: <br /><p><img style="display: block\; margin-left: auto\; margin-right: auto\;" src="https://events.vtools.ieee.org/vtools_ui /media/display/7f9b78fe-feb7-49fb-98ca-3e38fdf3fb93" alt="" width="556" he ight="371"></p>\n<p class="mb-4">Have you ever tried Google&rsquo\;s Live View feature\, where you just point your phone at a building and it instan tly figures out exactly where you are? Or maybe you&rsquo\;ve wondered how action movies pull off those cool 3D scenes of the places they&rsquo\;re about to rob? Ever stopped to think about how virtual tours and street vie ws come to life?</p>\n<p class="mb-4">In this talk\, we will explore the v arious algorithms that are used to create vivid and comprehensive 3D scene s from just a handful of images collected from the internet. The talk will be divided in 3 key sections:</p>\n<p class="mb-4"><strong>Artifact Mitig ation</strong><br>Images collected from the internet are not always perfec t. Some may have memes or text overlaid on them\, while others may have be en compressed multiple times for efficient storage\, resulting in pixelati on. Some images might be blurry\, have too much sun exposure\, or be taken at night. There can also be moving objects like people and cars\, which a ren&rsquo\;t needed and could obstruct the 3D reconstruction of a scene.</ p>\n<p class="mb-4">In this section\, we&rsquo\;ll learn about the deep le arning algorithms used to remove these kinds of artifacts and transient ob jects from images.</p>\n<p class="mb-4"><strong>Image Registration and Geo -localization</strong><br>Once the images have been pre-processed\, the ne xt step is to determine their relative pose with respect to each other. Im agine trying to figure out whether a photo was taken from the left side of the Eiffel Tower or the right\, or whether the person with the camera was 100 meters away or just 50 meters. Sometimes\, the images might even come from a drone! So how do we place all these different frames into a common reference frame? The more viewpoints we have of an area\, the more comple te our 3D models will be.</p>\n<p class="mb-4">In this section\, we&rsquo\ ;ll learn how Structure-from-Motion (SfM) is used to assign poses to these images. We&rsquo\;ll explore techniques for using background details to d etermine pose\, especially when the object of interest looks the same from every angle. And finally\, we&rsquo\;ll briefly discuss how these images can be geo-localized\; meaning their latitude and longitude can be estimat ed\, even when no GPS information is available.</p>\n<p class="mb-4"><stro ng>3D Reconstruction</strong><br>Now that our images are cleaned up and we know their poses\, we can dive into the techniques used to transform thes e images to build a 3D scene. We&rsquo\;ll discuss\, at a high level\, som e traditional 3D reconstruction methods along with more recent AI-based ap proaches such as neural rendering and gaussian splatting.</p>\n<p class="m b-4">The applications of these computer vision and deep learning technique s we&rsquo\;ve talked about are widespread\, and I hope you&rsquo\;ll star t noticing them being used all around you\, whether it&rsquo\;s in mobile phones\, self-driving technology\, biometric scanners\, delivery robots\, security cameras\, and many other places.</p><br /><br />Agenda: <br /><p> 7:00 PM &nbsp\; Networking and Announcements</p>\n<p>7:10 PM &nbsp\; Prese ntation</p> END:VEVENT END:VCALENDAR