/vlfeat_benchmarks

Benchmarking code for various computer vision algorithms

Primary LanguageMATLAB

This package implements detector benchmarks for various publically available
interest point detectors.

1. FIRST TIME INSTALLATION INSTRUCTIONS
Run the following commands from the matlab prompt. Make sure your present working
directory (pwd) is at the root (i.e. pwd looks like .../vlfeat_benchmarks/)

compile_mex(); % To compile the mex files, you need to have the mex compiler
               % setup properly for this to work
install();     % To install all third party dependencies

2. RUNNING THE BENCHMARK

setup();                         % Setups the paths properly
affineDetectors.benchmarkDemo(); % Runs a demo of the benchmark


3. CLEANING UP AN INSTALLATION

In case the install() command above is interrupted, or doesn't complete for some
reason, you can clear up all the downloaded third party software and re-install
from scratch by running the following:

clean();         % Will delete all the third party directories
install();       % To install all third party dependencies

4. MORE INFO

See help affineDetectors.benchmarkDemo for the list of currently available
detectors and datasets.

See help affineDetectors.runBenchmark for help on the actualy benchmarking
function and the options supported by it. affineDetectors.benchmarkDemo() calls
affineDetectors.runBenchmark() with some demo detectors.