Code Monkey home page Code Monkey logo

glimpse-project's Introduction

Glimpse, a General Layer-wise IMage ProceSsing Engine
=====================================================

The Glimpse project [IJCNN2013]_ is a library for implementing hierarchical
visual models in C++ and Python. The goal of this project is to allow a
broad range of feed-forward, hierarchical models to be encoded in a
high-level declarative manner, with low-level details of the implementation
hidden from view. This project combines an efficient implementation with the
ability to leverage parallel processing facilities and is designed to run on
multiple operating systems using only common, freely-available components. A
prototype of Glimpse has been used to encode an HMAX-like model, achieving
results comparable with those found in the literature. The project has been
supported by NSF Grant 1018967 (PIs: Melanie Mitchell and Garrett Kenyon).

Source code_ and documentation_ are available for the project, as well as an
installable package_.

.. _code: http://github.com/mthomure/glimpse-project
.. _documentation: http://pythonhosted.org/glimpse/
.. _package: http://pypi.python.org/pypi/glimpse

.. [IJCNN2013] Michael D. Thomure, Melanie Mitchell, Garrett T. Kenyon
   (2013). On the Role of Shape Prototypes in Hierarchical Models of Vision.
   To appear in *The International Joint Conference on Neural Networks
   (IJCNN)*.

glimpse-project's People

Contributors

mthomure avatar

Stargazers

avatar avatar avatar avatar avatar avatar avatar avatar avatar

Watchers

avatar avatar avatar avatar

Forkers

mandaarp nhanho tingfod georgedittmar seansouthern callistabee

glimpse-project's Issues

Version numbers on dependencies

Hello Mick,
I'm trying to set up Glimpse and I'm running into some problems getting the dependencies to work together. Do you think you could show a conda list of a working environment?
Thanks,
Sean Southern

Add better cross-validation behavior.

When cross-validation is used from GLAB, prototypes should optionally get re-learned for each sub-split. Currently, the prototypes are learned once, and then evaluation happens.

To support this, we'd need to refactor the ExperimentData object to combine the extractor and evaluation data. Alternatively, we could have cross-validation return multiple ExperimentData objects, where each corresponds to a different sub-split.

ML model ignores s1_shift_orientations.

The ml.model.s1_kernels accessor always shifts orientations, ignoring the s1_shift_orientations parameter. Update line 86 of glimpse/models/ml/model.py.

Since this affects past experiments, we should also update the default value of params.s1_shift_orientations.

Normalized dot-product at S1 should be absolute-value of NDP.

When the ML model uses an NDP activation function at S1, it's actually applying an absolute-value of NDP. This should be made more explicit by adding an ANDP (absolute-value NDP) operation to the backends, and removing the custom code at line 172 of glimpse/models/ml/model.py.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.