Introduction

X-Boost is a complete package performing Object Detection. The primary goal for this tool is fast train object detection and apply on image. This tool can also be used to provide benchmark between different object detection algorithms and techniques.

Installation

This tool requires cmake (or better ccmake) and build-essential application for your Linux OS.

Step 1: Compile the project

To compile and install the software create a build directory, go inside and execute ccmake:

mkdir build
cd build
ccmake .. 

Set CMAKE_BUILD_TYPE, the build mode, to Release, configure [c] and generate [g]. Finally compile the project

make

Step 2: Install X-Boost

Actually X-Boost is not ready to be installed. Use the applications in the build folder for now. In the feature application, library and development header will be installed.

Training

trainer is the tool used to train the classifier. See README.

Evaluating Performances

classify or roc can be used to evaluate the classifier performance. See README.

Use X-Boost as a Library

X-Boost package can be used as library inside external project. Include file and the xboost shared object must be installed and available in the path. After that you can use the example in src folder to create your classifier.

The classifier is allocated using allocate_classifier or allocate_detector functions.

See also: allocate_classifier allocate_detector