2AFC Sensitivity Statistic
Given a vector x of scoring values produced by a pattern classifier and a vector y of binary classification values, it computes the
sensitivity of the pattern classifier and the difficulty of each pattern. The sensitivity is measured in terms of he probability of being correct on a 2 alternative forced choice task. This measure of sensitivity is equivalent to the Area Under the ROC. (Matlab Function)
Click Here to download.
Multinomial Logistic Regression
A Matlab function implementing the Multinomial Logistic Regression classifier (see Tutorials Section for more information).
Click Here to download.
Nick’s Machine Perception Toolbox (NMPT)
Nick’s Machine Perception Toolkit (NMPT) is a C++ and OpenCV based open source Machine Perception API. It is being developed as part of an NSF funded project to use information maximization as a conceptual framework for unifying perception and action. The central philosophy of this package is three-fold: (1) Easy to Install; (2) Easy to Learn API; (3) Easy to Run. In this first release of Version 0.1, the main component of NMPT is an ultra-efficient implementation of the SUN model, a maximum-information approach to visual saliency developed by Zhang & Cottrell. Future versions of NMPT will also include facilities for optimal visual search. NMPT has been tested in MacOSX 10.4+ and Windows XP with Cygwin, and was developed to be fully cross-platform. Click Here to download.
RUBIOS 2.0 is a Java based Open Source communications API for for Social Robots. It was developed as part of the RUBI project to be the backbone for interconnecting the different perceptual and control modules of the RUBI robot. RUBI is a social robot developed to interact with toddlers and assist teachers in early childhood education settings. RUBIOS is has been battle tested in MacOSX 10.4.X and Windows XP with Cygwin. Click Here to download.
MPT Machine Perception Toolbox
Link to MPT Web Site
MPT supplies cross-platform C++ libraries for real-time perception primitives, including face detection, eye detection, blink detection, color tracking. Soon it will also include expression recognition, predictive color tracking, and tracking based on multisensor fusion. In addition, MPT supplies many examples which show how to embed the MPT funct ions in applications, and which are useful applications in their own right. Finally, Matlab .mex interfaces are provided for many of these core libraries for those who like using Matlab.
SCORE digital video coding tool
Developed by Joel Chenu, Click here to downlod.
Score is a toolbox we developed to facilitate manual coding of video sequences. It is a modular tool that can be easily modified for other video annotating applications. If you develop new Score tools and would like to make them available as part of the score package, please contact us at MPLAB
SCORE’s present form is a collection of utilities listed below:
Simple HMM Simulator written in C (targz archive)