BEGIN:VCALENDAR
VERSION:2.0
PRODID:www.dresden-science-calendar.de
METHOD:PUBLISH
CALSCALE:GREGORIAN
X-MICROSOFT-CALSCALE:GREGORIAN
X-WR-TIMEZONE:Europe/Berlin
BEGIN:VTIMEZONE
TZID:Europe/Berlin
X-LIC-LOCATION:Europe/Berlin
BEGIN:DAYLIGHT
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
DTSTART:19810329T030000
RRULE:FREQ=YEARLY;INTERVAL=1;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
DTSTART:19961027T030000
RRULE:FREQ=YEARLY;INTERVAL=1;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:DSC-13940
DTSTART;TZID=Europe/Berlin:20180117T100000
SEQUENCE:1516174901
TRANSP:OPAQUE
DTEND;TZID=Europe/Berlin:20180117T110000
URL:https://dresden-science-calendar.de/calendar/de/detail/13940
LOCATION:TUD Andreas-Pfitzmann-Bau\, Nöthnitzer Straße 4601069 Dresden
SUMMARY:Brachmann: Learning to Predict Dense Correspondences For 6D Pose Es
 timation
CLASS:PUBLIC
DESCRIPTION:Speaker: Dipl.-Medieninf. Eric Brachmann\nInstitute of Speaker:
  Institut für Software und Multimediatechnik\; Professur für Computergra
 phik und Visualisierung\nTopics:\nInformatik\n Location:\n  Name: TUD Andr
 eas-Pfitzmann-Bau (APB 1004 (Ratssaal))\n  Street: Nöthnitzer Straße 46\
 n  City: 01069 Dresden\n  Phone: \n  Fax: \nDescription: Object pose estim
 ation is an important problem in computer vision with applications in robo
 tics\, augmented reality and many other areas. An established strategy for
  object pose estimation consists of\, firstly\, finding correspondences be
 tween the image and the object's reference frame\, and\, secondly\, estima
 ting the pose from outlier-free correspondences using Random Sample Consen
 sus (RANSAC). The first step\, namely finding correspondences\, is difficu
 lt because object appearance varies depending on perspective\, lighting an
 d many other factors. Traditionally\, correspondences have been establishe
 d using handcrafted methods like sparse feature pipelines. In this thesis\
 , we introduce a dense correspondence representation for objects\, called 
 object coordinates\, which can be learned. By learning object coordinates\
 , our pose estimation pipeline adapts to various aspects of the task at ha
 nd. It works well for diverse object types\, from small objects to entire 
 rooms\, varying object attributes\, like textured or texture-less objects\
 , and different input modalities\, like RGB-D or RGB images. The concept o
 f object coordinates allows us to easily model and exploit uncertainty as 
 part of the pipeline such that even repeating structures or areas with lit
 tle texture can contribute to a good solution. Although we can train objec
 t coordinate predictors independent of the full pipeline and achieve good 
 results\, training the pipeline in an end-to-end fashion is desirable. It 
 enables the object coordinate predictor to adapt its output to the specifi
 cities of following steps in the pose estimation pipeline. Unfortunately\,
  the RANSAC component of the pipeline is non-differentiable which prohibit
 s end-to-end training. Adopting techniques from reinforcement learning\, w
 e introduce Differentiable Sample Consensus (DSAC)\, a formulation of RANS
 AC which allows us to train the pose estimation pipeline in an end-to-end 
 fashion by minimizing the expectation of the final pose error.
DTSTAMP:20260406T145729Z
CREATED:20180104T080156Z
LAST-MODIFIED:20180117T074141Z
END:VEVENT
END:VCALENDAR