pyleargist GIST Image descriptor for scene recognition
Download

pyleargist Ranking & Summary

pyleargist Tags

image descriptor Python library scene image descriptor scene recognition

pyleargist Description

GIST Image descriptor for scene recognition pyleargist is a Python library to compute GIST global image descriptors to be used to compare pictures based on their content (to be used global scene recognition and categorization).The GIST image descriptor theoritical definition can be found on A. Torralba's page: http://people.csail.mit.edu/torralba/code/spatialenvelope/The source code of the C implementation is included in the lear_gist subfolder. See http://lear.inrialpes.fr/software for the original project information.pyleargist is licensed under the GPL, the same license as the original C project.Installation:Install libfftw3 with development headers (http://www.fftw.org), python dev headers, gcc, the Python Imaging Library (PIL) and numpy.Build locally for testing:% python setup.py buid_ext -i% export PYTHONPATH=`pwd`/srcBuild and install system wide:% python setup.py build% sudo python setup.py installUsage:Here is a sample session in a python shell once the library is installed:>>> from PIL import Image>>> import leargist>>> im = Image.open('lear_gist/ar.ppm')>>> descriptors = leargist.color_gist(im)>>> descriptors.shape(960,)>>> descriptors.dtypedtype('float32')>>> descriptorsarray(, dtype=float32)The GIST descriptors (fixed size, 960 by default) can then be used as an euclidian space to cluster images based on their content.This dimension can then be reduced to a 32 or 64 bits semantic hash by using Locality Sensitive Hashing, Spectral Hashing or Stacked Denoising Autoencoders.A sample implementation of picture semantic hashing with SDAs is showcased in the libsgd library: http://code.oliviergrisel.name/libsgd Requirements: · Python What's New in This Release: · added missing missing MANIFEST

pyleargist Related Software