[dcchairs2016] UbiComp/ISWC 2016 Doctoral School notification - #121

[dcchairs2016 at ubicomp.org]

Dear Laura Cabrera-Quirós,

Please find enclosed the review for your accepted submission for the Ubicomp/ISWC 2016 Doctoral School.

121: "Towards multimodal analysis of human behavior in crowded mingling scenarios using movement cues from wearable sensors and cameras"

Please read the reviews and instructions that the committee members included in it carefully, as they contains necessary further steps to ensure your DC paper submission appears in the program. Any of the committee instructions has to be considered as required for your final version.

As a reminder, the final version ("camera ready" version) is due no later than Wednesday, July 17th, 2016 at 11:59pm PDT. We will send more instructions about how to upload the camera-ready version of your paper before the deadline.

Thank you for submitting your work to UbiComp 2016, and congratulations again on your acceptance to the Ubicomp 2016 DS. We look forward to seeing you in Heidelberg in September!


Max Mühlhäuser
Nadir Weibel
Rene Mayrhofer

UbiComp 2016 Doctoral School Chairs


------------------------ Submission 121, Review 1 ------------------------

Title: Towards multimodal analysis of human behavior in crowded mingling scenarios using movement cues from wearable sensors and cameras


Confidence

   3  (Very confident - I am knowledgeable in the area)

Contribution to UbiComp

   The paper is somewhat borderline w.r.t. UbiComp; both image analysis and
   worn sensors shall be used for understanding crowd events.

Overall Rating

   5  (Probably accept: I would argue for accepting this paper.)

The Review

   The paper aims at analyzing crowd events (parties, speed dating events,
   etc.) with different modalities, in particular image analysis (video
   cameras) and worn devices (accelerometers and proximity detection
   support). The PhD research shall forward the state of the art w.r.t.
   combined use of these modalities for better accuracy in detecting
   participants’ activities (gestures etc.), positions/encounters etc. As
   an important partial issue, the automatic association of wearables to
   persons shall be achieved with the help of the video data. In comparison
   to paper #115 (from Gedik, Ekin on "Estimating the state of a crowd"),
   the overall aims and research organization are less clear. On the other
   hand, the ultimate goal of “understanding crowd behavior” seems more
   well understood and somewhat more realistic (than the UX goal of #115);
   moreover, the section “developed work” is quite well presented and
   gives a very concise overview. If both papers get accepted, the
   similarities and differences w.r.t. ML approaches should be highlighted. 

   *************************
   Camera-ready requirements:

   The final version should be improved. The sections on past Research are
   already elaborate and would just benefit from another general improvement
   cycle; the chapter on planned research, however, should be elaborated
   further: the intended contributions and the corresponding methodological
   approaches are not described in sufficient detail (e.g., "enjoyment and
   attraction" are just stated as interesting attributes, but Little is said
   about corresponding cues, methods, etc.); as a consequence, the amount of
   (past and future) work / scientific achievements etc. is hard to judge -
   this is, however, important Input for the DC panel for providing valuable
   Feedback.

   Please also get in contact with Gedik, Ekin to discuss distinct
   presentation at the DC.