Progress for week 37 (2015)

Fra Robin

Gå til: navigasjon, søk

Innhold

Student template (copy this for your entry)

Budget

  • Todo 1
  • Todo 2

Accounting

  • Done 1
  • Done 2

Per Wessel

Budget

  • Lese papers om tids-serie anaylse av kjemikal sensorer
  • Kontakte og møte med Tobias Dahl hos Sintef
  • Finne potensielle veiledere og eller folk med kunnskap innen luft monitorering og sensoranalyse

Accounting

  • Skal møte Tobias Dahl fra Sintef denne uken
  • Er i dialog med NORA (Nitrous Oxide Reaserch Alliance) på nmbu (har ekspertise i tids-serie analyse av sensor data) og forurensning.

Magnus Olden

Budget

  • Read papers on ensemble learning / stacking
  • Finish essay
  • Move to latex/lyx
  • Learn more on Azure Machine Learning and SciKit Learn

Accounting

  • Read papers
  • Almost finished essay (4000 words, but still a little incoherent)
  • Moved to lyx
  • No decision yet on AML and ScKL

Tomas Renaa

Budget

  • Read:
    • essays and master theses.
    • read more papers to gain get better grasp at direction of thesis.
  • Look into MOCAP at Robin
  • Migrate document to LaTeX
  • Finish essay

Accounting

  • Done 1.1, 1.2
  • Done 3

Vegard

Budget

  • Read example essays
  • Finish at least half of the essay
  • Read at least half of part two of Williams and Bizup

Accounting

  • Read one example essay, skimmed the rest
  • 6-7 pages of essay written
  • Read the expected pages

Eirik Sundet

Budget

  • Recreate the region-growing clustering algorithm for road detection, implemented on the Stanley vehicle, who won the DARPA challenge in 2005.
  • explore good features for this task

Accounting

  • I've managed to estimate a gaussian model of the road in a single image, and then estimating the distance between each pixel and the mean of the road model.
  • I've managed to estimate a gaussian mixture model from 30 road images, with corresponding ground-truth masks.
  • I've begun to test how well the Gray-Level Co-Occurrence Matrix (GLCM) performs texture classification, in terms of accuracy and speed.
  • I've begun to explore some recursive algorithms for region-growth, but have not yet implemented a fully functional one.
Personlige verktøy