CBCL
SOFTWARE:
- Model of Object
Recognition:
Over the years, CBCL has developed and
implemented several different software versions of the model
of object recognition. The most up-to-date description of the
model is "A Theory of Object Recognition: Computations and
Circuits in the Feedforward Path of the Ventral Stream in
Primate Visual Cortex"
[Serre, Kouh, Cadieu, Knoblich, Kreiman, and Poggio (2005)]. The
table below has the links to the downloadable software.
| Version | Circa | Paper |
|---|
(1) "Bypass route only" code
A pure MATLAB implementation of the four-layer (S1->C2b) "bypass" route. Less general than (2-4),
but probably easier to understand. Runs about 3x slower than (3).
| 2005
| Serre,
Wolf & Poggio (CVPR 2005)
Serre,
Wolf, Bileschi, Riesenhuber & Poggio (PAMI 2007)
|
(2) Code for experiments replicating human performance in
the animal/no-animal rapid scene categorization task.
| 2007
| Serre,
Oliva & Poggio (PNAS 2007)
Serre,
Kouh, Cadieu, Knoblich, Kreiman & Poggio (AI memo 2005)
|
(3)
FHLib - Multiscale Feature Hierarchy Library
The most general version of the model. Allows the definition of feedforward hierarchies having an arbitrary
number of levels. MATLAB interface with a C++ back-end for efficiency. Arbitrary filter kernels can
be defined, but this requires creating new C++ classes and recompiling.
| 2007
|
Mutch & Lowe (IJCV 2007)
|
(4)
Model of object recognition with canonical normalization operations
Like (3), allows feedforward hierarchies of arbitrary depth, and is also MATLAB with a C++ back-end.
Defining new filters is easier than under (3), but all filters must be of the form
y = sum(f .* (x .^ p)) / (c + sum(f .* (x .^ q)) .^ r). Different settings
of c, p, q, and r correspond to tuning, softmax, etc. Models run more slowly than under (3).
| 2006
|
Kouh & Poggio (Neural Comp. 2008)
Zoccolan et al. (J. Neurosci. 2007)
Cadieu, et al. (J. Neurophys. 2007)
|
(5) HMAX "Hierarchical Model and X"
DISCLAIMER: This model implementation is now obsolete and is
no longer distributed.
| 1999
| Riesenhuber
and Poggio (Nat. Neurosci. 1999)
|
- Software for
Support Vector Machines
-
Regularized Least-Squares MATLAB Toolkit
- WordEx: A word extraction
engine
CBCL DATASETS:
CBCL
SUPPLEMENTARY MATERIAL:
The SVMFu software and RLS Toolkit is released under
a standard open source license, but the CBCL data sets are
covered by the following copyright.
Copyright
2000
Center for Biological & Computational Learning at
MIT and MIT
All rights reserved.
Permission to copy and modify this data, software, and
its documentation only for internal research use in your
organization is hereby granted, provided that this notice
is retained thereon and on all copies. This data and software
should not be distributed to anyone outside of your organization
without explicit written authorization by the author(s)
and MIT. It should not be used for commercial purposes
without specific permission from the authors and MIT.
MIT also requires written authorization by the author(s)
to publish results obtained with the data or software
and possibly citation of relevant CBCL reference papers.
We make no representation as to the suitability and operability
of this data or software for any purpose. It is provided
"as is" without express or implied warranty.
|
