The STAIR Vision Library (v2.3)

This document provides a detailed description of the design and implementation of the STAIR Vision Library. It includes an overview of the various components of the library, installation instructions, coding conventions, and outline of specific classes within the library. The intended audience is users and developers of the STAIR Vision Library. The document assumes that you are familiar with either the Linux or Windows (Microsoft Visual Studio) development environments, and are an experienced C/C++ programmer. Since the code is under continuous development and improvement, the ultimate specification for classes within the library is the code itself. When using the library you should consult the various headers for specific details on expected arguments, etc.. Separate tutorial/HOWTO documents provide detail step-bystep instructions on the use of various applications and projects that are built using the library. The SVL is a library of research code initially developed to support the STanford AI Project (STAIR). It has since been expanded to provide a range of software infrastructure for computer vision, machine learning, and probabilistic graphical models. However, the library has maintained its focus as a vehicle for research and this is reflected in its design—many of the classes in the library are only loosely coupled allowing researchers to pull out only what they need and link against their code. The library is intended to be platform independent and currently compiles on both Linux and Windows. We have released the library under the BSD license (see Appendix C).