Olfaction

From Vision Wiki
Jump to navigation Jump to search

Log | People | Tasks | Daily Notebook

Projects

Smell the world

Collect statistics of the smells in the natural world.

Odor recognition

Recognize objects based on a sniff in a real-world environment. Compare with blindfolded human performance.

Olfaction + vision

As above when one has access to both an image and the odor.

Challenges

  • Sensor drift & calibration
  • Invariance to water vapor (with and without specific sensor)
  • Detect individual components of mixtures
  • Learning: when do we learn a new odor? How far can we go unsupervised?

Progress Reports

Here is my Daily Olfaction Notebook

Each week I'll make a brief progress summary below.

Documentation

Documents from Michael Gibson:

User Documentation for ENose Board.

Protocol for communicating with the board.

Data Processing needed to convert hex values to meaningful resistances.

Literature

  1. Brian Sisk's 2004 thesis PDF
  2. R Basri, T Hassner, L Zelnik-Manor - Computer Vision and Pattern Recognition, 2007. CVPR'07. IEEE …, 2007 PDF