Progress for week 47 (2014)

Fra Robin

(Forskjeller mellom versjoner)
Gå til: navigasjon, søk
(Else-Line)
(Eivind)
 
(Én mellomrevisjon ikke vist.)
Linje 17: Linje 17:
=== Accounting ===
=== Accounting ===
 +
* Changed evaluation project to x64 due to memory issues.
 +
* Ran further tests on SGBM, BM and ELAS
 +
* Dataset appears mostly OK, SGBM just slow with 256 disparity levels.
 +
* Discussed cameras with OmniVision.
 +
* Took another look at how current benchmarks compare algorithms and display their results.
 +
* Started experimenting with QT for easier visualization of results.
== Eivind ==
== Eivind ==
Linje 25: Linje 31:
=== Accounting ===
=== Accounting ===
 +
* Did performance tests and wrote the chapters about them
 +
* Painted the game area!
 +
* Wrote a couple of pages
== Else-Line ==
== Else-Line ==

Nåværende revisjon fra 25. nov 2014 kl. 10:00

Innhold

Student template (copy this for your entry)

Budget

  • Todo 1
  • Todo 2

Accounting

  • Done 1
  • Done 2

Stian

Budget

  • Continue moving algorithms into the framework.
    • rSGM or perhaps SPS?
  • Consider tweaking the new dataset. Large images makes for very slow progress.
  • Start comparing results
    • Look into good methods of doing so.

Accounting

  • Changed evaluation project to x64 due to memory issues.
  • Ran further tests on SGBM, BM and ELAS
  • Dataset appears mostly OK, SGBM just slow with 256 disparity levels.
  • Discussed cameras with OmniVision.
  • Took another look at how current benchmarks compare algorithms and display their results.
  • Started experimenting with QT for easier visualization of results.

Eivind

Budget

  • Performance-tests to aquire useful data
  • Build/paint game area!
  • Spend at least two days writing.

Accounting

  • Did performance tests and wrote the chapters about them
  • Painted the game area!
  • Wrote a couple of pages

Else-Line

Budget

  • Continue testing
    • Try different parameters for the local search (number of iterations, lambda)
    • Try different parameters for the evolution?
  • Start planning how the thesis should look

Accounting

  • Ran more tests
    • crossover vs no crossover
    • 5 vs 10 iterations
    • 20/200 vs 40/400 generations
  • Started displaying data in graphs
  • Did some small changes to the system (stop on iterations finished instead of time)

Rune

Budget

  • Finish the PRM-implementation
  • Improve the scenario generator (more robots).

Accounting

  • Probabilistic road map algorithm implemented

Snorre

Budget

  • Start to look into transferability measurements
  • Plan more experiments

Accounting

Bendik

Budget

  • Write 5 pages
  • Paint!

André

Budget

  • Paint
  • Continue the design of the boxes
  • Figure out what approach, if slip ring, or mirror and a stepper with hollow shaft.

Accounting

  • Painted the game area
  • Done with a prototype of the box with a gear. Have to buy a ordinary slip ring to make the tower be able to turn multiple rounds
  • Think the stepper works fine, and read out the ange correct.
Personlige verktøy