In situ depth maps based feature extraction and tracking

Yucong Chris Ye, Yang Wang, Robert Miller, Kwan-Liu Ma, Kenji Ono

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Scopus citations


Parallel numerical simulation is a powerful tool used by scientists to study complex problems. It has been a common practice to save the simulation output to disk and then conduct post-hoc in-depth analyses of the saved data. System I/O capabilities have not kept pace as simulations have scaled up over time, so a common approach has been to output only subsets of the data to reduce I/O. However, as we are entering the era of peta-and exa-scale computing, this sub-sampling approach is no longer acceptable because too much valuable information is lost. In situ visualization has been shown a promising approach to the data problem at extreme-scale. We present a novel in situ solution using depth maps to enable post-hoc image-based visualization and feature extraction and tracking. An interactive interface is provided to allow for fine-tuning the generation of depth maps during the course of a simulation run to better capture the features of interest. We use several applications including one actual simulation run on a Cray XE6 supercomputer to demonstrate the effectiveness of our approach.

Original languageEnglish (US)
Title of host publicationIEEE Symposium on Large Data Analysis and Visualization 2015, LDAV 2015 - Proceedings
EditorsJanine Bennett, Hank Childs, Markus Hadwiger, Hank Childs
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages8
ISBN (Electronic)9781467385176
StatePublished - Dec 4 2015
Event5th IEEE Symposium on Large Data Analysis and Visualization, LDAV 2015 - Chicago, United States
Duration: Oct 25 2015Oct 26 2015


Other5th IEEE Symposium on Large Data Analysis and Visualization, LDAV 2015
Country/TerritoryUnited States

ASJC Scopus subject areas

  • Communication
  • Computer Networks and Communications
  • Signal Processing


Dive into the research topics of 'In situ depth maps based feature extraction and tracking'. Together they form a unique fingerprint.

Cite this