Computer Vision Platform
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
apps
bobo
data
matlab
test
INSTALL.txt
README.md

README.md

Project: BOBO - Computer Vision Platform
Author: Jeffrey Byrne jeff@visym.com
URL: https://github.com/visym/bobo/

Overview

The BOBO project is a set of matlab and python client tools for distributed stream processing of computer vision tasks. These tools are designed to interface with a Spark cluster for performance analysis of large datasets.

Design Goals

  • BOBO is designed for interactive spark based distributed processing of large computer vision datasets
  • BOBO is designed for researchers performing analysis of large image datasets
  • BOBO provides a tools for typical computer vision performance evaluations (object recognition and detection)
  • BOBO is designed for easy deployment of vision tasks on a spark production cluster
  • BOBO is open source and written in python and C++, and designed for iPython interactive usage

Supported Datasets

  1. ImageNet-Fall2011
  2. Caltech101
  3. Caltech256
  4. ETHZ shapes
  5. ETHZ extended shapes
  6. MNIST
  7. LabelMe3
  8. Pascal VOC 2012
  9. PubFig
  10. Labeled Faces in the Wild (LFW)
  11. LFW - Deep Funneled
  12. LFW - Funneled

Dependencies

Required: numpy, opencv, matplotlib, sklearn, h5py, pylab
Optional: flickrapi, nltk, pygame

Dataset Dependencies

nltk.download() wordnet corpus