Software developed in ISFER
ISFER framework:
The framework for the entire facial features recognition system. Implemented in Java by Maarten van Schouwen and Jos Vollering. General information about the workbench can be found in the publication "A multimedia workbench for facial expression analysis". More detailed technical information can be found in internal tech-reports and on the ISFER programmers page .Modules available for ISFER framework:
General image processing modules:- Image2Color - Converts a java picture to the Colour data type.
- Color2Grey - Converts a Colour picture to a grey-scale picture.
- ImageRotator - Rotates a java picture over given angle.
- Grey2Image - Converts Grey data to the java picture format.
- Convolution - Use the mask to do a convolution of the input picture.
- Median - Take the median of the input picture.
- DrawPolygon - Draws the polygon in the grey-scale picture.
Framework image processing modules:
- Grey2MRP - Converts the Grey data to the Multi Resolution Pyramid.
- GreyClip - Clips the features from the grey-scale image.
Emotion recognition modules:
- MRP2RFM - Finds global features.
- FindHeadContour - Locates the head contour in the input picture.
- FindNoseChin - Locates the nose and chin in the input picture.
- FindMouthArea - Finds the exact mouth feature map.
- SnakeMouth - Locates mouth contour using active contours.
- CurvefitMouth - Fits a curve around the mouth.
- FuzzyMouth - Calculates an average edge direction in some areas of the picture.
- EdgeData2Grey - Visualises the edge direction data.
- FuzzyNN - A neural network classifying the edge direction data.
- NNEye - A neural network finding the microfeatures for the eye
- SnakeEye - Locates an eye contour using active contours.
- FindEyebrow - Locates eyebrow in the input picture.
- SimplifyPolygon - Simplifies the eyebrow polygon to just a few points.
- SetFeatureMap - An interactive module that allows the user to manually set the Features Maps.
- SetFCP - An interactive module that allows the user to manually set the Facial Characteristic Points