# HG changeset patch
# User Robert McIntyre <rlm@mit.edu>
# Date 1361896289 0
# Node ID eb7c94a0318819c5f521edf8dc2b3ef8dacf22d3
# Parent  e5636b1740f834887e26849ac22c9c7f645a7665
thesis notes.

diff -r e5636b1740f8 -r eb7c94a03188 org/notes.org
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/org/notes.org	Tue Feb 26 16:31:29 2013 +0000
@@ -0,0 +1,50 @@
+
+
+* scribbles 
+
+scratch.mit.edu
+
+fredkin 
+ef@cmu.edu
+
+gary Drescher logo papert lab
+
+* stuff for thesis from Winston 2/22
+
+Shimon Ullman tank -- a 3D model vision thing
+
+Paul Viola MIT PHD thesis (intensity -> surface)
+
+Grimson -- multiple scales
+
+Winston -- "most effective way to find papers is to ask someone."
+
+Talk to Poggio/Ullman/Finlayson to get moar papers
+
+
+* Notes from Vision meeting Wednesday, 2/20
+
+David Clemems, PHG Thesis
+Arizona university mind's eye project used model based vision
+
+"Model based vision" -> "generative vision"
+
+Microsoft Research flickr city square
+
+celiu -- M$ motion guy 
+
+prakesh -- read paper
+
+from ADK:
+
+I wasn't able to find much in an hour but I probably just don't know
+the magic keywords.  You should look at David Clemens's thesis as I
+mentioned. Also I believe it is Paul Cohen from Arizona State
+University who initially worked on the simulation-driven activity
+recognition in Mind's Eye, and then afaik changed tack completely.
+
+http://w3.sista.arizona.edu/~cohen/Publications/
+
+There is also Michael Burl of JPL whose page seems unavailable right
+now. The JPL team also used 3D models of pose along with some kind of
+scripts to recognize events.