[Lisa_teatalk] Tea talk tomorrow Thu 13th Feb, 14h-15h in Z300

Razvan Pascanu r.pascanu at gmail.com
Tue Feb 11 08:28:19 EST 2014


Hi all,

Thursday we have a tea talk by Jyri Kivinen. The room is going to be Z300
McNicoll pavilion, time 14h00 to 15h00. PLEASE NOTE THE CHANGE OF TIME AND
PLACE.

Here is the title and abstract:

Title:
Statistical Models for Natural Scene Data

Abstract:
This talk describes work largely described in my PhD thesis (University of
Edinburgh, 2013) titled Statistical Models for Natural Scene Data.

We consider statistical modelling of natural image data. Obtaining advances
in this field can have significant impact for both engineering
applications, and for the understanding of the human visual system. Several
recent advances in natural image modeling have been obtained with the use
of unsupervised feature learning. We consider a class of such models,
restricted Boltzmann machines (RBMs), used in many recent state-of-the-art
image models. We develop extensions of these stochastic artificial neural
networks, and use them as a basis for building more effective image models,
and tools for computational vision.

First, motivated by the unsatisfactory results of current generative
natural image models, we take a step back, and evaluate whether they are
able to model a subclass of the data, natural image textures. This is a
necessary subcomponent of any credible model for visual scenes. We assess
the performance of a state-of-the-art model of natural images for texture
generation, using a dataset and evaluation techniques from in prior work.
We also perform a dissection of the model architecture, uncovering the
properties important for good performance.  Building on this, we develop
structured extensions for more complicated data comprised of textures from
multiple classes, using the single-texture
model architecture as a basis. These models are shown to be able to produce
state-of-the-art texture synthesis results quantitatively, and are also
effective qualitatively. It is demonstrated empirically that the developed
multiple-texture framework provides a means to generate images of
differently textured regions, more generic globally varying textures, and
can also be used for texture interpolation, where the approach is radically
different from the others in the area.

We then consider visual boundary prediction from natural images. The work
aims to improve understanding of Boltzmann machines in the generation of
image segment boundaries, and to investigate deep neural network
architectures for learning the boundary detection problem. The developed
networks (which avoid several hand-crafted model and feature designs
commonly used for the problem), produce effective inference with
state-of-the-art or better performance.

Finally we discuss a novel framework we have developed for obtaining
Boltzmann machines, in which the hidden unit activations co-transform with
transformed input stimuli in a stable and predictable way throughout the
network. We define such models to be \emph{transformation equivariant}.
Such properties have been shown useful for computer vision systems, and
have been motivational for example in the development of steerable filters,
a widely used classical feature extraction technique. Translation
equivariant feature sharing has been the standard method for scaling image
models beyond patch-sized data to large images. In our framework we extend
shallow and deep models to account for other kinds of transformations as
well, focusing on in-plane rotations.

I hope to see many of you there !

Best,
Razvan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://webmail.iro.umontreal.ca/pipermail/lisa_teatalk/attachments/20140211/35c2947b/attachment.html 


More information about the Lisa_teatalk mailing list