[6/3] High Performance Analytics

0 意見

[6/3] Support vector learning for ordinal regression

0 意見
pros:
  classification methodologies can be directly applied
  training data are easy to get
cons:
  minimize error in classification rather than ranking
  pairwise computation cost

[5/27] Geometric min-Hashing: Finding a (Thick) Needle in a Haystack

0 意見

[5/20] Fast concurrent object localization and recognition

0 意見

[5/13] Learning object categories from Google's image search

0 意見
training image from google, testing on Caltech dataset

unsupervised learning algorithm

pLSA
  no spatial information
ABS-pLSA
  latent variable x (grid centroid coordinate)
  grid methodology, still translation and scaling variant
TSI-pLSA
  latent variable c:(centroid, xscale, yscale)
  1. apply to pLSA
  2. fit a mixture of Gaussian with k=(1, 2, ...K) components to the location of the region, weighted by p(w|z)
  3.

[4/29] Object class recognition by unsupervised scale-invariant learning

0 意見

[4/22] Photo Tourism: Exploring Photo Collection in 3D

0 意見